ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1396 - (show annotations)
Sun Nov 10 19:04:34 2013 UTC (5 years, 11 months ago) by ph10
File size: 133386 byte(s)
In /x mode, allow white space before a possessive + character.
1 .TH PCREPATTERN 3 "09 November 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 (*UCP).
84 This has the same effect as setting the PCRE_UCP option: it causes sequences
85 such as \ed and \ew to use Unicode properties to determine character types,
86 instead of recognizing only characters with codes less than 128 via a lookup
87 table.
88 .
89 .
90 .SS "Disabling auto-possessification"
91 .rs
92 .sp
93 If a pattern starts with (*NO_AUTO_POSSESS), it has the same effect as setting
94 the PCRE_NO_AUTO_POSSESSIFY option at compile time. This stops PCRE from making
95 quantifiers possessive when what follows cannot match the repeated item. For
96 example, by default a+b is treated as a++b. For more details, see the
97 .\" HREF
98 \fBpcreapi\fP
99 .\"
100 documentation.
101 .
102 .
103 .SS "Disabling start-up optimizations"
104 .rs
105 .sp
106 If a pattern starts with (*NO_START_OPT), it has the same effect as setting the
107 PCRE_NO_START_OPTIMIZE option either at compile or matching time. This disables
108 several optimizations for quickly reaching "no match" results. For more
109 details, see the
110 .\" HREF
111 \fBpcreapi\fP
112 .\"
113 documentation.
114 .
115 .
116 .\" HTML <a name="newlines"></a>
117 .SS "Newline conventions"
118 .rs
119 .sp
120 PCRE supports five different conventions for indicating line breaks in
121 strings: a single CR (carriage return) character, a single LF (linefeed)
122 character, the two-character sequence CRLF, any of the three preceding, or any
123 Unicode newline sequence. The
124 .\" HREF
125 \fBpcreapi\fP
126 .\"
127 page has
128 .\" HTML <a href="pcreapi.html#newlines">
129 .\" </a>
130 further discussion
131 .\"
132 about newlines, and shows how to set the newline convention in the
133 \fIoptions\fP arguments for the compiling and matching functions.
134 .P
135 It is also possible to specify a newline convention by starting a pattern
136 string with one of the following five sequences:
137 .sp
138 (*CR) carriage return
139 (*LF) linefeed
140 (*CRLF) carriage return, followed by linefeed
141 (*ANYCRLF) any of the three above
142 (*ANY) all Unicode newline sequences
143 .sp
144 These override the default and the options given to the compiling function. For
145 example, on a Unix system where LF is the default newline sequence, the pattern
146 .sp
147 (*CR)a.b
148 .sp
149 changes the convention to CR. That pattern matches "a\enb" because LF is no
150 longer a newline. If more than one of these settings is present, the last one
151 is used.
152 .P
153 The newline convention affects where the circumflex and dollar assertions are
154 true. It also affects the interpretation of the dot metacharacter when
155 PCRE_DOTALL is not set, and the behaviour of \eN. However, it does not affect
156 what the \eR escape sequence matches. By default, this is any Unicode newline
157 sequence, for Perl compatibility. However, this can be changed; see the
158 description of \eR in the section entitled
159 .\" HTML <a href="#newlineseq">
160 .\" </a>
161 "Newline sequences"
162 .\"
163 below. A change of \eR setting can be combined with a change of newline
164 convention.
165 .
166 .
167 .SS "Setting match and recursion limits"
168 .rs
169 .sp
170 The caller of \fBpcre_exec()\fP can set a limit on the number of times the
171 internal \fBmatch()\fP function is called and on the maximum depth of
172 recursive calls. These facilities are provided to catch runaway matches that
173 are provoked by patterns with huge matching trees (a typical example is a
174 pattern with nested unlimited repeats) and to avoid running out of system stack
175 by too much recursion. When one of these limits is reached, \fBpcre_exec()\fP
176 gives an error return. The limits can also be set by items at the start of the
177 pattern of the form
178 .sp
179 (*LIMIT_MATCH=d)
181 .sp
182 where d is any number of decimal digits. However, the value of the setting must
183 be less than the value set (or defaulted) by the caller of \fBpcre_exec()\fP
184 for it to have any effect. In other words, the pattern writer can lower the
185 limits set by the programmer, but not raise them. If there is more than one
186 setting of one of these limits, the lower value is used.
187 .
188 .
190 .rs
191 .sp
192 PCRE can be compiled to run in an environment that uses EBCDIC as its character
193 code rather than ASCII or Unicode (typically a mainframe system). In the
194 sections below, character code values are ASCII or Unicode; in an EBCDIC
195 environment these characters may have different code values, and there are no
196 code points greater than 255.
197 .
198 .
200 .rs
201 .sp
202 A regular expression is a pattern that is matched against a subject string from
203 left to right. Most characters stand for themselves in a pattern, and match the
204 corresponding characters in the subject. As a trivial example, the pattern
205 .sp
206 The quick brown fox
207 .sp
208 matches a portion of a subject string that is identical to itself. When
209 caseless matching is specified (the PCRE_CASELESS option), letters are matched
210 independently of case. In a UTF mode, PCRE always understands the concept of
211 case for characters whose values are less than 128, so caseless matching is
212 always possible. For characters with higher values, the concept of case is
213 supported if PCRE is compiled with Unicode property support, but not otherwise.
214 If you want to use caseless matching for characters 128 and above, you must
215 ensure that PCRE is compiled with Unicode property support as well as with
216 UTF support.
217 .P
218 The power of regular expressions comes from the ability to include alternatives
219 and repetitions in the pattern. These are encoded in the pattern by the use of
220 \fImetacharacters\fP, which do not stand for themselves but instead are
221 interpreted in some special way.
222 .P
223 There are two different sets of metacharacters: those that are recognized
224 anywhere in the pattern except within square brackets, and those that are
225 recognized within square brackets. Outside square brackets, the metacharacters
226 are as follows:
227 .sp
228 \e general escape character with several uses
229 ^ assert start of string (or line, in multiline mode)
230 $ assert end of string (or line, in multiline mode)
231 . match any character except newline (by default)
232 [ start character class definition
233 | start of alternative branch
234 ( start subpattern
235 ) end subpattern
236 ? extends the meaning of (
237 also 0 or 1 quantifier
238 also quantifier minimizer
239 * 0 or more quantifier
240 + 1 or more quantifier
241 also "possessive quantifier"
242 { start min/max quantifier
243 .sp
244 Part of a pattern that is in square brackets is called a "character class". In
245 a character class the only metacharacters are:
246 .sp
247 \e general escape character
248 ^ negate the class, but only if the first character
249 - indicates character range
250 .\" JOIN
251 [ POSIX character class (only if followed by POSIX
252 syntax)
253 ] terminates the character class
254 .sp
255 The following sections describe the use of each of the metacharacters.
256 .
257 .
259 .rs
260 .sp
261 The backslash character has several uses. Firstly, if it is followed by a
262 character that is not a number or a letter, it takes away any special meaning
263 that character may have. This use of backslash as an escape character applies
264 both inside and outside character classes.
265 .P
266 For example, if you want to match a * character, you write \e* in the pattern.
267 This escaping action applies whether or not the following character would
268 otherwise be interpreted as a metacharacter, so it is always safe to precede a
269 non-alphanumeric with backslash to specify that it stands for itself. In
270 particular, if you want to match a backslash, you write \e\e.
271 .P
272 In a UTF mode, only ASCII numbers and letters have any special meaning after a
273 backslash. All other characters (in particular, those whose codepoints are
274 greater than 127) are treated as literals.
275 .P
276 If a pattern is compiled with the PCRE_EXTENDED option, most white space in the
277 pattern (other than in a character class), and characters between a # outside a
278 character class and the next newline, inclusive, are ignored. An escaping
279 backslash can be used to include a white space or # character as part of the
280 pattern.
281 .P
282 If you want to remove the special meaning from a sequence of characters, you
283 can do so by putting them between \eQ and \eE. This is different from Perl in
284 that $ and @ are handled as literals in \eQ...\eE sequences in PCRE, whereas in
285 Perl, $ and @ cause variable interpolation. Note the following examples:
286 .sp
287 Pattern PCRE matches Perl matches
288 .sp
289 .\" JOIN
290 \eQabc$xyz\eE abc$xyz abc followed by the
291 contents of $xyz
292 \eQabc\e$xyz\eE abc\e$xyz abc\e$xyz
293 \eQabc\eE\e$\eQxyz\eE abc$xyz abc$xyz
294 .sp
295 The \eQ...\eE sequence is recognized both inside and outside character classes.
296 An isolated \eE that is not preceded by \eQ is ignored. If \eQ is not followed
297 by \eE later in the pattern, the literal interpretation continues to the end of
298 the pattern (that is, \eE is assumed at the end). If the isolated \eQ is inside
299 a character class, this causes an error, because the character class is not
300 terminated.
301 .
302 .
303 .\" HTML <a name="digitsafterbackslash"></a>
304 .SS "Non-printing characters"
305 .rs
306 .sp
307 A second use of backslash provides a way of encoding non-printing characters
308 in patterns in a visible manner. There is no restriction on the appearance of
309 non-printing characters, apart from the binary zero that terminates a pattern,
310 but when a pattern is being prepared by text editing, it is often easier to use
311 one of the following escape sequences than the binary character it represents:
312 .sp
313 \ea alarm, that is, the BEL character (hex 07)
314 \ecx "control-x", where x is any ASCII character
315 \ee escape (hex 1B)
316 \ef form feed (hex 0C)
317 \en linefeed (hex 0A)
318 \er carriage return (hex 0D)
319 \et tab (hex 09)
320 \e0dd character with octal code 0dd
321 \eddd character with octal code ddd, or back reference
322 \eo{ddd..} character with octal code ddd..
323 \exhh character with hex code hh
324 \ex{hhh..} character with hex code hhh.. (non-JavaScript mode)
325 \euhhhh character with hex code hhhh (JavaScript mode only)
326 .sp
327 The precise effect of \ecx on ASCII characters is as follows: if x is a lower
328 case letter, it is converted to upper case. Then bit 6 of the character (hex
329 40) is inverted. Thus \ecA to \ecZ become hex 01 to hex 1A (A is 41, Z is 5A),
330 but \ec{ becomes hex 3B ({ is 7B), and \ec; becomes hex 7B (; is 3B). If the
331 data item (byte or 16-bit value) following \ec has a value greater than 127, a
332 compile-time error occurs. This locks out non-ASCII characters in all modes.
333 .P
334 The \ec facility was designed for use with ASCII characters, but with the
335 extension to Unicode it is even less useful than it once was. It is, however,
336 recognized when PCRE is compiled in EBCDIC mode, where data items are always
337 bytes. In this mode, all values are valid after \ec. If the next character is a
338 lower case letter, it is converted to upper case. Then the 0xc0 bits of the
339 byte are inverted. Thus \ecA becomes hex 01, as in ASCII (A is C1), but because
340 the EBCDIC letters are disjoint, \ecZ becomes hex 29 (Z is E9), and other
341 characters also generate different values.
342 .P
343 After \e0 up to two further octal digits are read. If there are fewer than two
344 digits, just those that are present are used. Thus the sequence \e0\ex\e07
345 specifies two binary zeros followed by a BEL character (code value 7). Make
346 sure you supply two digits after the initial zero if the pattern character that
347 follows is itself an octal digit.
348 .P
349 The escape \eo must be followed by a sequence of octal digits, enclosed in
350 braces. An error occurs if this is not the case. This escape is a recent
351 addition to Perl; it provides way of specifying character code points as octal
352 numbers greater than 0777, and it also allows octal numbers and back references
353 to be unambiguously specified.
354 .P
355 For greater clarity and unambiguity, it is best to avoid following \e by a
356 digit greater than zero. Instead, use \eo{} or \ex{} to specify character
357 numbers, and \eg{} to specify back references. The following paragraphs
358 describe the old, ambiguous syntax.
359 .P
360 The handling of a backslash followed by a digit other than 0 is complicated,
361 and Perl has changed in recent releases, causing PCRE also to change. Outside a
362 character class, PCRE reads the digit and any following digits as a decimal
363 number. If the number is less than 8, or if there have been at least that many
364 previous capturing left parentheses in the expression, the entire sequence is
365 taken as a \fIback reference\fP. A description of how this works is given
366 .\" HTML <a href="#backreferences">
367 .\" </a>
368 later,
369 .\"
370 following the discussion of
371 .\" HTML <a href="#subpattern">
372 .\" </a>
373 parenthesized subpatterns.
374 .\"
375 .P
376 Inside a character class, or if the decimal number following \e is greater than
377 7 and there have not been that many capturing subpatterns, PCRE handles \e8 and
378 \e9 as the literal characters "8" and "9", and otherwise re-reads up to three
379 octal digits following the backslash, using them to generate a data character.
380 Any subsequent digits stand for themselves. For example:
381 .sp
382 \e040 is another way of writing an ASCII space
383 .\" JOIN
384 \e40 is the same, provided there are fewer than 40
385 previous capturing subpatterns
386 \e7 is always a back reference
387 .\" JOIN
388 \e11 might be a back reference, or another way of
389 writing a tab
390 \e011 is always a tab
391 \e0113 is a tab followed by the character "3"
392 .\" JOIN
393 \e113 might be a back reference, otherwise the
394 character with octal code 113
395 .\" JOIN
396 \e377 might be a back reference, otherwise
397 the value 255 (decimal)
398 .\" JOIN
399 \e81 is either a back reference, or the two
400 characters "8" and "1"
401 .sp
402 Note that octal values of 100 or greater that are specified using this syntax
403 must not be introduced by a leading zero, because no more than three octal
404 digits are ever read.
405 .P
406 By default, after \ex that is not followed by {, from zero to two hexadecimal
407 digits are read (letters can be in upper or lower case). Any number of
408 hexadecimal digits may appear between \ex{ and }. If a character other than
409 a hexadecimal digit appears between \ex{ and }, or if there is no terminating
410 }, an error occurs.
411 .P
412 If the PCRE_JAVASCRIPT_COMPAT option is set, the interpretation of \ex is
413 as just described only when it is followed by two hexadecimal digits.
414 Otherwise, it matches a literal "x" character. In JavaScript mode, support for
415 code points greater than 256 is provided by \eu, which must be followed by
416 four hexadecimal digits; otherwise it matches a literal "u" character.
417 .P
418 Characters whose value is less than 256 can be defined by either of the two
419 syntaxes for \ex (or by \eu in JavaScript mode). There is no difference in the
420 way they are handled. For example, \exdc is exactly the same as \ex{dc} (or
421 \eu00dc in JavaScript mode).
422 .
423 .
424 .SS "Constraints on character values"
425 .rs
426 .sp
427 Characters that are specified using octal or hexadecimal numbers are
428 limited to certain values, as follows:
429 .sp
430 8-bit non-UTF mode less than 0x100
431 8-bit UTF-8 mode less than 0x10ffff and a valid codepoint
432 16-bit non-UTF mode less than 0x10000
433 16-bit UTF-16 mode less than 0x10ffff and a valid codepoint
434 32-bit non-UTF mode less than 0x100000000
435 32-bit UTF-32 mode less than 0x10ffff and a valid codepoint
436 .sp
437 Invalid Unicode codepoints are the range 0xd800 to 0xdfff (the so-called
438 "surrogate" codepoints), and 0xffef.
439 .
440 .
441 .SS "Escape sequences in character classes"
442 .rs
443 .sp
444 All the sequences that define a single character value can be used both inside
445 and outside character classes. In addition, inside a character class, \eb is
446 interpreted as the backspace character (hex 08).
447 .P
448 \eN is not allowed in a character class. \eB, \eR, and \eX are not special
449 inside a character class. Like other unrecognized escape sequences, they are
450 treated as the literal characters "B", "R", and "X" by default, but cause an
451 error if the PCRE_EXTRA option is set. Outside a character class, these
452 sequences have different meanings.
453 .
454 .
455 .SS "Unsupported escape sequences"
456 .rs
457 .sp
458 In Perl, the sequences \el, \eL, \eu, and \eU are recognized by its string
459 handler and used to modify the case of following characters. By default, PCRE
460 does not support these escape sequences. However, if the PCRE_JAVASCRIPT_COMPAT
461 option is set, \eU matches a "U" character, and \eu can be used to define a
462 character by code point, as described in the previous section.
463 .
464 .
465 .SS "Absolute and relative back references"
466 .rs
467 .sp
468 The sequence \eg followed by an unsigned or a negative number, optionally
469 enclosed in braces, is an absolute or relative back reference. A named back
470 reference can be coded as \eg{name}. Back references are discussed
471 .\" HTML <a href="#backreferences">
472 .\" </a>
473 later,
474 .\"
475 following the discussion of
476 .\" HTML <a href="#subpattern">
477 .\" </a>
478 parenthesized subpatterns.
479 .\"
480 .
481 .
482 .SS "Absolute and relative subroutine calls"
483 .rs
484 .sp
485 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
486 a number enclosed either in angle brackets or single quotes, is an alternative
487 syntax for referencing a subpattern as a "subroutine". Details are discussed
488 .\" HTML <a href="#onigurumasubroutines">
489 .\" </a>
490 later.
491 .\"
492 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
493 synonymous. The former is a back reference; the latter is a
494 .\" HTML <a href="#subpatternsassubroutines">
495 .\" </a>
496 subroutine
497 .\"
498 call.
499 .
500 .
501 .\" HTML <a name="genericchartypes"></a>
502 .SS "Generic character types"
503 .rs
504 .sp
505 Another use of backslash is for specifying generic character types:
506 .sp
507 \ed any decimal digit
508 \eD any character that is not a decimal digit
509 \eh any horizontal white space character
510 \eH any character that is not a horizontal white space character
511 \es any white space character
512 \eS any character that is not a white space character
513 \ev any vertical white space character
514 \eV any character that is not a vertical white space character
515 \ew any "word" character
516 \eW any "non-word" character
517 .sp
518 There is also the single sequence \eN, which matches a non-newline character.
519 This is the same as
520 .\" HTML <a href="#fullstopdot">
521 .\" </a>
522 the "." metacharacter
523 .\"
524 when PCRE_DOTALL is not set. Perl also uses \eN to match characters by name;
525 PCRE does not support this.
526 .P
527 Each pair of lower and upper case escape sequences partitions the complete set
528 of characters into two disjoint sets. Any given character matches one, and only
529 one, of each pair. The sequences can appear both inside and outside character
530 classes. They each match one character of the appropriate type. If the current
531 matching point is at the end of the subject string, all of them fail, because
532 there is no character to match.
533 .P
534 For compatibility with Perl, \es did not used to match the VT character (code
535 11), which made it different from the the POSIX "space" class. However, Perl
536 added VT at release 5.18, and PCRE followed suit at release 8.34. The \es
537 characters are now HT (9), LF (10), VT (11), FF (12), CR (13), and space (32).
538 .P
539 A "word" character is an underscore or any character that is a letter or digit.
540 By default, the definition of letters and digits is controlled by PCRE's
541 low-valued character tables, and may vary if locale-specific matching is taking
542 place (see
543 .\" HTML <a href="pcreapi.html#localesupport">
544 .\" </a>
545 "Locale support"
546 .\"
547 in the
548 .\" HREF
549 \fBpcreapi\fP
550 .\"
551 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
552 or "french" in Windows, some character codes greater than 128 are used for
553 accented letters, and these are then matched by \ew. The use of locales with
554 Unicode is discouraged.
555 .P
556 By default, in a UTF mode, characters with values greater than 128 never match
557 \ed, \es, or \ew, and always match \eD, \eS, and \eW. These sequences retain
558 their original meanings from before UTF support was available, mainly for
559 efficiency reasons. However, if PCRE is compiled with Unicode property support,
560 and the PCRE_UCP option is set, the behaviour is changed so that Unicode
561 properties are used to determine character types, as follows:
562 .sp
563 \ed any character that matches \ep{Nd} (decimal digit)
564 \es any character that matches \ep{Z} or \eh or \ev
565 \ew any character that matches \ep{L} or \ep{N}, plus underscore
566 .sp
567 The upper case escapes match the inverse sets of characters. Note that \ed
568 matches only decimal digits, whereas \ew matches any Unicode digit, as well as
569 any Unicode letter, and underscore. Note also that PCRE_UCP affects \eb, and
570 \eB because they are defined in terms of \ew and \eW. Matching these sequences
571 is noticeably slower when PCRE_UCP is set.
572 .P
573 The sequences \eh, \eH, \ev, and \eV are features that were added to Perl at
574 release 5.10. In contrast to the other sequences, which match only ASCII
575 characters by default, these always match certain high-valued codepoints,
576 whether or not PCRE_UCP is set. The horizontal space characters are:
577 .sp
578 U+0009 Horizontal tab (HT)
579 U+0020 Space
580 U+00A0 Non-break space
581 U+1680 Ogham space mark
582 U+180E Mongolian vowel separator
583 U+2000 En quad
584 U+2001 Em quad
585 U+2002 En space
586 U+2003 Em space
587 U+2004 Three-per-em space
588 U+2005 Four-per-em space
589 U+2006 Six-per-em space
590 U+2007 Figure space
591 U+2008 Punctuation space
592 U+2009 Thin space
593 U+200A Hair space
594 U+202F Narrow no-break space
595 U+205F Medium mathematical space
596 U+3000 Ideographic space
597 .sp
598 The vertical space characters are:
599 .sp
600 U+000A Linefeed (LF)
601 U+000B Vertical tab (VT)
602 U+000C Form feed (FF)
603 U+000D Carriage return (CR)
604 U+0085 Next line (NEL)
605 U+2028 Line separator
606 U+2029 Paragraph separator
607 .sp
608 In 8-bit, non-UTF-8 mode, only the characters with codepoints less than 256 are
609 relevant.
610 .
611 .
612 .\" HTML <a name="newlineseq"></a>
613 .SS "Newline sequences"
614 .rs
615 .sp
616 Outside a character class, by default, the escape sequence \eR matches any
617 Unicode newline sequence. In 8-bit non-UTF-8 mode \eR is equivalent to the
618 following:
619 .sp
620 (?>\er\en|\en|\ex0b|\ef|\er|\ex85)
621 .sp
622 This is an example of an "atomic group", details of which are given
623 .\" HTML <a href="#atomicgroup">
624 .\" </a>
625 below.
626 .\"
627 This particular group matches either the two-character sequence CR followed by
628 LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
629 U+000B), FF (form feed, U+000C), CR (carriage return, U+000D), or NEL (next
630 line, U+0085). The two-character sequence is treated as a single unit that
631 cannot be split.
632 .P
633 In other modes, two additional characters whose codepoints are greater than 255
634 are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
635 Unicode character property support is not needed for these characters to be
636 recognized.
637 .P
638 It is possible to restrict \eR to match only CR, LF, or CRLF (instead of the
639 complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
640 either at compile time or when the pattern is matched. (BSR is an abbrevation
641 for "backslash R".) This can be made the default when PCRE is built; if this is
642 the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
643 It is also possible to specify these settings by starting a pattern string with
644 one of the following sequences:
645 .sp
646 (*BSR_ANYCRLF) CR, LF, or CRLF only
647 (*BSR_UNICODE) any Unicode newline sequence
648 .sp
649 These override the default and the options given to the compiling function, but
650 they can themselves be overridden by options given to a matching function. Note
651 that these special settings, which are not Perl-compatible, are recognized only
652 at the very start of a pattern, and that they must be in upper case. If more
653 than one of them is present, the last one is used. They can be combined with a
654 change of newline convention; for example, a pattern can start with:
655 .sp
657 .sp
658 They can also be combined with the (*UTF8), (*UTF16), (*UTF32), (*UTF) or
659 (*UCP) special sequences. Inside a character class, \eR is treated as an
660 unrecognized escape sequence, and so matches the letter "R" by default, but
661 causes an error if PCRE_EXTRA is set.
662 .
663 .
664 .\" HTML <a name="uniextseq"></a>
665 .SS Unicode character properties
666 .rs
667 .sp
668 When PCRE is built with Unicode character property support, three additional
669 escape sequences that match characters with specific properties are available.
670 When in 8-bit non-UTF-8 mode, these sequences are of course limited to testing
671 characters whose codepoints are less than 256, but they do work in this mode.
672 The extra escape sequences are:
673 .sp
674 \ep{\fIxx\fP} a character with the \fIxx\fP property
675 \eP{\fIxx\fP} a character without the \fIxx\fP property
676 \eX a Unicode extended grapheme cluster
677 .sp
678 The property names represented by \fIxx\fP above are limited to the Unicode
679 script names, the general category properties, "Any", which matches any
680 character (including newline), and some special PCRE properties (described
681 in the
682 .\" HTML <a href="#extraprops">
683 .\" </a>
684 next section).
685 .\"
686 Other Perl properties such as "InMusicalSymbols" are not currently supported by
687 PCRE. Note that \eP{Any} does not match any characters, so always causes a
688 match failure.
689 .P
690 Sets of Unicode characters are defined as belonging to certain scripts. A
691 character from one of these sets can be matched using a script name. For
692 example:
693 .sp
694 \ep{Greek}
695 \eP{Han}
696 .sp
697 Those that are not part of an identified script are lumped together as
698 "Common". The current list of scripts is:
699 .P
700 Arabic,
701 Armenian,
702 Avestan,
703 Balinese,
704 Bamum,
705 Batak,
706 Bengali,
707 Bopomofo,
708 Brahmi,
709 Braille,
710 Buginese,
711 Buhid,
712 Canadian_Aboriginal,
713 Carian,
714 Chakma,
715 Cham,
716 Cherokee,
717 Common,
718 Coptic,
719 Cuneiform,
720 Cypriot,
721 Cyrillic,
722 Deseret,
723 Devanagari,
724 Egyptian_Hieroglyphs,
725 Ethiopic,
726 Georgian,
727 Glagolitic,
728 Gothic,
729 Greek,
730 Gujarati,
731 Gurmukhi,
732 Han,
733 Hangul,
734 Hanunoo,
735 Hebrew,
736 Hiragana,
737 Imperial_Aramaic,
738 Inherited,
739 Inscriptional_Pahlavi,
740 Inscriptional_Parthian,
741 Javanese,
742 Kaithi,
743 Kannada,
744 Katakana,
745 Kayah_Li,
746 Kharoshthi,
747 Khmer,
748 Lao,
749 Latin,
750 Lepcha,
751 Limbu,
752 Linear_B,
753 Lisu,
754 Lycian,
755 Lydian,
756 Malayalam,
757 Mandaic,
758 Meetei_Mayek,
759 Meroitic_Cursive,
760 Meroitic_Hieroglyphs,
761 Miao,
762 Mongolian,
763 Myanmar,
764 New_Tai_Lue,
765 Nko,
766 Ogham,
767 Old_Italic,
768 Old_Persian,
769 Old_South_Arabian,
770 Old_Turkic,
771 Ol_Chiki,
772 Oriya,
773 Osmanya,
774 Phags_Pa,
775 Phoenician,
776 Rejang,
777 Runic,
778 Samaritan,
779 Saurashtra,
780 Sharada,
781 Shavian,
782 Sinhala,
783 Sora_Sompeng,
784 Sundanese,
785 Syloti_Nagri,
786 Syriac,
787 Tagalog,
788 Tagbanwa,
789 Tai_Le,
790 Tai_Tham,
791 Tai_Viet,
792 Takri,
793 Tamil,
794 Telugu,
795 Thaana,
796 Thai,
797 Tibetan,
798 Tifinagh,
799 Ugaritic,
800 Vai,
801 Yi.
802 .P
803 Each character has exactly one Unicode general category property, specified by
804 a two-letter abbreviation. For compatibility with Perl, negation can be
805 specified by including a circumflex between the opening brace and the property
806 name. For example, \ep{^Lu} is the same as \eP{Lu}.
807 .P
808 If only one letter is specified with \ep or \eP, it includes all the general
809 category properties that start with that letter. In this case, in the absence
810 of negation, the curly brackets in the escape sequence are optional; these two
811 examples have the same effect:
812 .sp
813 \ep{L}
814 \epL
815 .sp
816 The following general category property codes are supported:
817 .sp
818 C Other
819 Cc Control
820 Cf Format
821 Cn Unassigned
822 Co Private use
823 Cs Surrogate
824 .sp
825 L Letter
826 Ll Lower case letter
827 Lm Modifier letter
828 Lo Other letter
829 Lt Title case letter
830 Lu Upper case letter
831 .sp
832 M Mark
833 Mc Spacing mark
834 Me Enclosing mark
835 Mn Non-spacing mark
836 .sp
837 N Number
838 Nd Decimal number
839 Nl Letter number
840 No Other number
841 .sp
842 P Punctuation
843 Pc Connector punctuation
844 Pd Dash punctuation
845 Pe Close punctuation
846 Pf Final punctuation
847 Pi Initial punctuation
848 Po Other punctuation
849 Ps Open punctuation
850 .sp
851 S Symbol
852 Sc Currency symbol
853 Sk Modifier symbol
854 Sm Mathematical symbol
855 So Other symbol
856 .sp
857 Z Separator
858 Zl Line separator
859 Zp Paragraph separator
860 Zs Space separator
861 .sp
862 The special property L& is also supported: it matches a character that has
863 the Lu, Ll, or Lt property, in other words, a letter that is not classified as
864 a modifier or "other".
865 .P
866 The Cs (Surrogate) property applies only to characters in the range U+D800 to
867 U+DFFF. Such characters are not valid in Unicode strings and so
868 cannot be tested by PCRE, unless UTF validity checking has been turned off
869 (see the discussion of PCRE_NO_UTF8_CHECK, PCRE_NO_UTF16_CHECK and
870 PCRE_NO_UTF32_CHECK in the
871 .\" HREF
872 \fBpcreapi\fP
873 .\"
874 page). Perl does not support the Cs property.
875 .P
876 The long synonyms for property names that Perl supports (such as \ep{Letter})
877 are not supported by PCRE, nor is it permitted to prefix any of these
878 properties with "Is".
879 .P
880 No character that is in the Unicode table has the Cn (unassigned) property.
881 Instead, this property is assumed for any code point that is not in the
882 Unicode table.
883 .P
884 Specifying caseless matching does not affect these escape sequences. For
885 example, \ep{Lu} always matches only upper case letters. This is different from
886 the behaviour of current versions of Perl.
887 .P
888 Matching characters by Unicode property is not fast, because PCRE has to do a
889 multistage table lookup in order to find a character's property. That is why
890 the traditional escape sequences such as \ed and \ew do not use Unicode
891 properties in PCRE by default, though you can make them do so by setting the
892 PCRE_UCP option or by starting the pattern with (*UCP).
893 .
894 .
895 .SS Extended grapheme clusters
896 .rs
897 .sp
898 The \eX escape matches any number of Unicode characters that form an "extended
899 grapheme cluster", and treats the sequence as an atomic group
900 .\" HTML <a href="#atomicgroup">
901 .\" </a>
902 (see below).
903 .\"
904 Up to and including release 8.31, PCRE matched an earlier, simpler definition
905 that was equivalent to
906 .sp
907 (?>\ePM\epM*)
908 .sp
909 That is, it matched a character without the "mark" property, followed by zero
910 or more characters with the "mark" property. Characters with the "mark"
911 property are typically non-spacing accents that affect the preceding character.
912 .P
913 This simple definition was extended in Unicode to include more complicated
914 kinds of composite character by giving each character a grapheme breaking
915 property, and creating rules that use these properties to define the boundaries
916 of extended grapheme clusters. In releases of PCRE later than 8.31, \eX matches
917 one of these clusters.
918 .P
919 \eX always matches at least one character. Then it decides whether to add
920 additional characters according to the following rules for ending a cluster:
921 .P
922 1. End at the end of the subject string.
923 .P
924 2. Do not end between CR and LF; otherwise end after any control character.
925 .P
926 3. Do not break Hangul (a Korean script) syllable sequences. Hangul characters
927 are of five types: L, V, T, LV, and LVT. An L character may be followed by an
928 L, V, LV, or LVT character; an LV or V character may be followed by a V or T
929 character; an LVT or T character may be follwed only by a T character.
930 .P
931 4. Do not end before extending characters or spacing marks. Characters with
932 the "mark" property always have the "extend" grapheme breaking property.
933 .P
934 5. Do not end after prepend characters.
935 .P
936 6. Otherwise, end the cluster.
937 .
938 .
939 .\" HTML <a name="extraprops"></a>
940 .SS PCRE's additional properties
941 .rs
942 .sp
943 As well as the standard Unicode properties described above, PCRE supports four
944 more that make it possible to convert traditional escape sequences such as \ew
945 and \es to use Unicode properties. PCRE uses these non-standard, non-Perl
946 properties internally when PCRE_UCP is set. However, they may also be used
947 explicitly. These properties are:
948 .sp
949 Xan Any alphanumeric character
950 Xps Any POSIX space character
951 Xsp Any Perl space character
952 Xwd Any Perl "word" character
953 .sp
954 Xan matches characters that have either the L (letter) or the N (number)
955 property. Xps matches the characters tab, linefeed, vertical tab, form feed, or
956 carriage return, and any other character that has the Z (separator) property.
957 Xsp is the same as Xps; it used to exclude vertical tab, for Perl
958 compatibility, but Perl changed, and so PCRE followed at release 8.34. Xwd
959 matches the same characters as Xan, plus underscore.
960 .P
961 There is another non-standard property, Xuc, which matches any character that
962 can be represented by a Universal Character Name in C++ and other programming
963 languages. These are the characters $, @, ` (grave accent), and all characters
964 with Unicode code points greater than or equal to U+00A0, except for the
965 surrogates U+D800 to U+DFFF. Note that most base (ASCII) characters are
966 excluded. (Universal Character Names are of the form \euHHHH or \eUHHHHHHHH
967 where H is a hexadecimal digit. Note that the Xuc property does not match these
968 sequences but the characters that they represent.)
969 .
970 .
971 .\" HTML <a name="resetmatchstart"></a>
972 .SS "Resetting the match start"
973 .rs
974 .sp
975 The escape sequence \eK causes any previously matched characters not to be
976 included in the final matched sequence. For example, the pattern:
977 .sp
978 foo\eKbar
979 .sp
980 matches "foobar", but reports that it has matched "bar". This feature is
981 similar to a lookbehind assertion
982 .\" HTML <a href="#lookbehind">
983 .\" </a>
984 (described below).
985 .\"
986 However, in this case, the part of the subject before the real match does not
987 have to be of fixed length, as lookbehind assertions do. The use of \eK does
988 not interfere with the setting of
989 .\" HTML <a href="#subpattern">
990 .\" </a>
991 captured substrings.
992 .\"
993 For example, when the pattern
994 .sp
995 (foo)\eKbar
996 .sp
997 matches "foobar", the first substring is still set to "foo".
998 .P
999 Perl documents that the use of \eK within assertions is "not well defined". In
1000 PCRE, \eK is acted upon when it occurs inside positive assertions, but is
1001 ignored in negative assertions.
1002 .
1003 .
1004 .\" HTML <a name="smallassertions"></a>
1005 .SS "Simple assertions"
1006 .rs
1007 .sp
1008 The final use of backslash is for certain simple assertions. An assertion
1009 specifies a condition that has to be met at a particular point in a match,
1010 without consuming any characters from the subject string. The use of
1011 subpatterns for more complicated assertions is described
1012 .\" HTML <a href="#bigassertions">
1013 .\" </a>
1014 below.
1015 .\"
1016 The backslashed assertions are:
1017 .sp
1018 \eb matches at a word boundary
1019 \eB matches when not at a word boundary
1020 \eA matches at the start of the subject
1021 \eZ matches at the end of the subject
1022 also matches before a newline at the end of the subject
1023 \ez matches only at the end of the subject
1024 \eG matches at the first matching position in the subject
1025 .sp
1026 Inside a character class, \eb has a different meaning; it matches the backspace
1027 character. If any other of these assertions appears in a character class, by
1028 default it matches the corresponding literal character (for example, \eB
1029 matches the letter B). However, if the PCRE_EXTRA option is set, an "invalid
1030 escape sequence" error is generated instead.
1031 .P
1032 A word boundary is a position in the subject string where the current character
1033 and the previous character do not both match \ew or \eW (i.e. one matches
1034 \ew and the other matches \eW), or the start or end of the string if the
1035 first or last character matches \ew, respectively. In a UTF mode, the meanings
1036 of \ew and \eW can be changed by setting the PCRE_UCP option. When this is
1037 done, it also affects \eb and \eB. Neither PCRE nor Perl has a separate "start
1038 of word" or "end of word" metasequence. However, whatever follows \eb normally
1039 determines which it is. For example, the fragment \eba matches "a" at the start
1040 of a word.
1041 .P
1042 The \eA, \eZ, and \ez assertions differ from the traditional circumflex and
1043 dollar (described in the next section) in that they only ever match at the very
1044 start and end of the subject string, whatever options are set. Thus, they are
1045 independent of multiline mode. These three assertions are not affected by the
1046 PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
1047 circumflex and dollar metacharacters. However, if the \fIstartoffset\fP
1048 argument of \fBpcre_exec()\fP is non-zero, indicating that matching is to start
1049 at a point other than the beginning of the subject, \eA can never match. The
1050 difference between \eZ and \ez is that \eZ matches before a newline at the end
1051 of the string as well as at the very end, whereas \ez matches only at the end.
1052 .P
1053 The \eG assertion is true only when the current matching position is at the
1054 start point of the match, as specified by the \fIstartoffset\fP argument of
1055 \fBpcre_exec()\fP. It differs from \eA when the value of \fIstartoffset\fP is
1056 non-zero. By calling \fBpcre_exec()\fP multiple times with appropriate
1057 arguments, you can mimic Perl's /g option, and it is in this kind of
1058 implementation where \eG can be useful.
1059 .P
1060 Note, however, that PCRE's interpretation of \eG, as the start of the current
1061 match, is subtly different from Perl's, which defines it as the end of the
1062 previous match. In Perl, these can be different when the previously matched
1063 string was empty. Because PCRE does just one match at a time, it cannot
1064 reproduce this behaviour.
1065 .P
1066 If all the alternatives of a pattern begin with \eG, the expression is anchored
1067 to the starting match position, and the "anchored" flag is set in the compiled
1068 regular expression.
1069 .
1070 .
1072 .rs
1073 .sp
1074 The circumflex and dollar metacharacters are zero-width assertions. That is,
1075 they test for a particular condition being true without consuming any
1076 characters from the subject string.
1077 .P
1078 Outside a character class, in the default matching mode, the circumflex
1079 character is an assertion that is true only if the current matching point is at
1080 the start of the subject string. If the \fIstartoffset\fP argument of
1081 \fBpcre_exec()\fP is non-zero, circumflex can never match if the PCRE_MULTILINE
1082 option is unset. Inside a character class, circumflex has an entirely different
1083 meaning
1084 .\" HTML <a href="#characterclass">
1085 .\" </a>
1086 (see below).
1087 .\"
1088 .P
1089 Circumflex need not be the first character of the pattern if a number of
1090 alternatives are involved, but it should be the first thing in each alternative
1091 in which it appears if the pattern is ever to match that branch. If all
1092 possible alternatives start with a circumflex, that is, if the pattern is
1093 constrained to match only at the start of the subject, it is said to be an
1094 "anchored" pattern. (There are also other constructs that can cause a pattern
1095 to be anchored.)
1096 .P
1097 The dollar character is an assertion that is true only if the current matching
1098 point is at the end of the subject string, or immediately before a newline at
1099 the end of the string (by default). Note, however, that it does not actually
1100 match the newline. Dollar need not be the last character of the pattern if a
1101 number of alternatives are involved, but it should be the last item in any
1102 branch in which it appears. Dollar has no special meaning in a character class.
1103 .P
1104 The meaning of dollar can be changed so that it matches only at the very end of
1105 the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
1106 does not affect the \eZ assertion.
1107 .P
1108 The meanings of the circumflex and dollar characters are changed if the
1109 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
1110 immediately after internal newlines as well as at the start of the subject
1111 string. It does not match after a newline that ends the string. A dollar
1112 matches before any newlines in the string, as well as at the very end, when
1113 PCRE_MULTILINE is set. When newline is specified as the two-character
1114 sequence CRLF, isolated CR and LF characters do not indicate newlines.
1115 .P
1116 For example, the pattern /^abc$/ matches the subject string "def\enabc" (where
1117 \en represents a newline) in multiline mode, but not otherwise. Consequently,
1118 patterns that are anchored in single line mode because all branches start with
1119 ^ are not anchored in multiline mode, and a match for circumflex is possible
1120 when the \fIstartoffset\fP argument of \fBpcre_exec()\fP is non-zero. The
1121 PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
1122 .P
1123 Note that the sequences \eA, \eZ, and \ez can be used to match the start and
1124 end of the subject in both modes, and if all branches of a pattern start with
1125 \eA it is always anchored, whether or not PCRE_MULTILINE is set.
1126 .
1127 .
1128 .\" HTML <a name="fullstopdot"></a>
1130 .rs
1131 .sp
1132 Outside a character class, a dot in the pattern matches any one character in
1133 the subject string except (by default) a character that signifies the end of a
1134 line.
1135 .P
1136 When a line ending is defined as a single character, dot never matches that
1137 character; when the two-character sequence CRLF is used, dot does not match CR
1138 if it is immediately followed by LF, but otherwise it matches all characters
1139 (including isolated CRs and LFs). When any Unicode line endings are being
1140 recognized, dot does not match CR or LF or any of the other line ending
1141 characters.
1142 .P
1143 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
1144 option is set, a dot matches any one character, without exception. If the
1145 two-character sequence CRLF is present in the subject string, it takes two dots
1146 to match it.
1147 .P
1148 The handling of dot is entirely independent of the handling of circumflex and
1149 dollar, the only relationship being that they both involve newlines. Dot has no
1150 special meaning in a character class.
1151 .P
1152 The escape sequence \eN behaves like a dot, except that it is not affected by
1153 the PCRE_DOTALL option. In other words, it matches any character except one
1154 that signifies the end of a line. Perl also uses \eN to match characters by
1155 name; PCRE does not support this.
1156 .
1157 .
1159 .rs
1160 .sp
1161 Outside a character class, the escape sequence \eC matches any one data unit,
1162 whether or not a UTF mode is set. In the 8-bit library, one data unit is one
1163 byte; in the 16-bit library it is a 16-bit unit; in the 32-bit library it is
1164 a 32-bit unit. Unlike a dot, \eC always
1165 matches line-ending characters. The feature is provided in Perl in order to
1166 match individual bytes in UTF-8 mode, but it is unclear how it can usefully be
1167 used. Because \eC breaks up characters into individual data units, matching one
1168 unit with \eC in a UTF mode means that the rest of the string may start with a
1169 malformed UTF character. This has undefined results, because PCRE assumes that
1170 it is dealing with valid UTF strings (and by default it checks this at the
1171 start of processing unless the PCRE_NO_UTF8_CHECK, PCRE_NO_UTF16_CHECK or
1172 PCRE_NO_UTF32_CHECK option is used).
1173 .P
1174 PCRE does not allow \eC to appear in lookbehind assertions
1175 .\" HTML <a href="#lookbehind">
1176 .\" </a>
1177 (described below)
1178 .\"
1179 in a UTF mode, because this would make it impossible to calculate the length of
1180 the lookbehind.
1181 .P
1182 In general, the \eC escape sequence is best avoided. However, one
1183 way of using it that avoids the problem of malformed UTF characters is to use a
1184 lookahead to check the length of the next character, as in this pattern, which
1185 could be used with a UTF-8 string (ignore white space and line breaks):
1186 .sp
1187 (?| (?=[\ex00-\ex7f])(\eC) |
1188 (?=[\ex80-\ex{7ff}])(\eC)(\eC) |
1189 (?=[\ex{800}-\ex{ffff}])(\eC)(\eC)(\eC) |
1190 (?=[\ex{10000}-\ex{1fffff}])(\eC)(\eC)(\eC)(\eC))
1191 .sp
1192 A group that starts with (?| resets the capturing parentheses numbers in each
1193 alternative (see
1194 .\" HTML <a href="#dupsubpatternnumber">
1195 .\" </a>
1196 "Duplicate Subpattern Numbers"
1197 .\"
1198 below). The assertions at the start of each branch check the next UTF-8
1199 character for values whose encoding uses 1, 2, 3, or 4 bytes, respectively. The
1200 character's individual bytes are then captured by the appropriate number of
1201 groups.
1202 .
1203 .
1204 .\" HTML <a name="characterclass"></a>
1206 .rs
1207 .sp
1208 An opening square bracket introduces a character class, terminated by a closing
1209 square bracket. A closing square bracket on its own is not special by default.
1210 However, if the PCRE_JAVASCRIPT_COMPAT option is set, a lone closing square
1211 bracket causes a compile-time error. If a closing square bracket is required as
1212 a member of the class, it should be the first data character in the class
1213 (after an initial circumflex, if present) or escaped with a backslash.
1214 .P
1215 A character class matches a single character in the subject. In a UTF mode, the
1216 character may be more than one data unit long. A matched character must be in
1217 the set of characters defined by the class, unless the first character in the
1218 class definition is a circumflex, in which case the subject character must not
1219 be in the set defined by the class. If a circumflex is actually required as a
1220 member of the class, ensure it is not the first character, or escape it with a
1221 backslash.
1222 .P
1223 For example, the character class [aeiou] matches any lower case vowel, while
1224 [^aeiou] matches any character that is not a lower case vowel. Note that a
1225 circumflex is just a convenient notation for specifying the characters that
1226 are in the class by enumerating those that are not. A class that starts with a
1227 circumflex is not an assertion; it still consumes a character from the subject
1228 string, and therefore it fails if the current pointer is at the end of the
1229 string.
1230 .P
1231 In UTF-8 (UTF-16, UTF-32) mode, characters with values greater than 255 (0xffff)
1232 can be included in a class as a literal string of data units, or by using the
1233 \ex{ escaping mechanism.
1234 .P
1235 When caseless matching is set, any letters in a class represent both their
1236 upper case and lower case versions, so for example, a caseless [aeiou] matches
1237 "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
1238 caseful version would. In a UTF mode, PCRE always understands the concept of
1239 case for characters whose values are less than 128, so caseless matching is
1240 always possible. For characters with higher values, the concept of case is
1241 supported if PCRE is compiled with Unicode property support, but not otherwise.
1242 If you want to use caseless matching in a UTF mode for characters 128 and
1243 above, you must ensure that PCRE is compiled with Unicode property support as
1244 well as with UTF support.
1245 .P
1246 Characters that might indicate line breaks are never treated in any special way
1247 when matching character classes, whatever line-ending sequence is in use, and
1248 whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
1249 such as [^a] always matches one of these characters.
1250 .P
1251 The minus (hyphen) character can be used to specify a range of characters in a
1252 character class. For example, [d-m] matches any letter between d and m,
1253 inclusive. If a minus character is required in a class, it must be escaped with
1254 a backslash or appear in a position where it cannot be interpreted as
1255 indicating a range, typically as the first or last character in the class, or
1256 immediately after a range. For example, [b-d-z] matches letters in the range b
1257 to d, a hyphen character, or z.
1258 .P
1259 It is not possible to have the literal character "]" as the end character of a
1260 range. A pattern such as [W-]46] is interpreted as a class of two characters
1261 ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
1262 "-46]". However, if the "]" is escaped with a backslash it is interpreted as
1263 the end of range, so [W-\e]46] is interpreted as a class containing a range
1264 followed by two other characters. The octal or hexadecimal representation of
1265 "]" can also be used to end a range.
1266 .P
1267 An error is generated if a POSIX character class (see below) or an escape
1268 sequence other than one that defines a single character appears at a point
1269 where a range ending character is expected. For example, [z-\exff] is valid,
1270 but [A-\ed] and [A-[:digit:]] are not.
1271 .P
1272 Ranges operate in the collating sequence of character values. They can also be
1273 used for characters specified numerically, for example [\e000-\e037]. Ranges
1274 can include any characters that are valid for the current mode.
1275 .P
1276 If a range that includes letters is used when caseless matching is set, it
1277 matches the letters in either case. For example, [W-c] is equivalent to
1278 [][\e\e^_`wxyzabc], matched caselessly, and in a non-UTF mode, if character
1279 tables for a French locale are in use, [\exc8-\excb] matches accented E
1280 characters in both cases. In UTF modes, PCRE supports the concept of case for
1281 characters with values greater than 128 only when it is compiled with Unicode
1282 property support.
1283 .P
1284 The character escape sequences \ed, \eD, \eh, \eH, \ep, \eP, \es, \eS, \ev,
1285 \eV, \ew, and \eW may appear in a character class, and add the characters that
1286 they match to the class. For example, [\edABCDEF] matches any hexadecimal
1287 digit. In UTF modes, the PCRE_UCP option affects the meanings of \ed, \es, \ew
1288 and their upper case partners, just as it does when they appear outside a
1289 character class, as described in the section entitled
1290 .\" HTML <a href="#genericchartypes">
1291 .\" </a>
1292 "Generic character types"
1293 .\"
1294 above. The escape sequence \eb has a different meaning inside a character
1295 class; it matches the backspace character. The sequences \eB, \eN, \eR, and \eX
1296 are not special inside a character class. Like any other unrecognized escape
1297 sequences, they are treated as the literal characters "B", "N", "R", and "X" by
1298 default, but cause an error if the PCRE_EXTRA option is set.
1299 .P
1300 A circumflex can conveniently be used with the upper case character types to
1301 specify a more restricted set of characters than the matching lower case type.
1302 For example, the class [^\eW_] matches any letter or digit, but not underscore,
1303 whereas [\ew] includes underscore. A positive character class should be read as
1304 "something OR something OR ..." and a negative class as "NOT something AND NOT
1305 something AND NOT ...".
1306 .P
1307 The only metacharacters that are recognized in character classes are backslash,
1308 hyphen (only where it can be interpreted as specifying a range), circumflex
1309 (only at the start), opening square bracket (only when it can be interpreted as
1310 introducing a POSIX class name - see the next section), and the terminating
1311 closing square bracket. However, escaping other non-alphanumeric characters
1312 does no harm.
1313 .
1314 .
1316 .rs
1317 .sp
1318 Perl supports the POSIX notation for character classes. This uses names
1319 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
1320 this notation. For example,
1321 .sp
1322 [01[:alpha:]%]
1323 .sp
1324 matches "0", "1", any alphabetic character, or "%". The supported class names
1325 are:
1326 .sp
1327 alnum letters and digits
1328 alpha letters
1329 ascii character codes 0 - 127
1330 blank space or tab only
1331 cntrl control characters
1332 digit decimal digits (same as \ed)
1333 graph printing characters, excluding space
1334 lower lower case letters
1335 print printing characters, including space
1336 punct printing characters, excluding letters and digits and space
1337 space white space (the same as \es from PCRE 8.34)
1338 upper upper case letters
1339 word "word" characters (same as \ew)
1340 xdigit hexadecimal digits
1341 .sp
1342 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
1343 space (32). "Space" used to be different to \es, which did not include VT, for
1344 Perl compatibility. However, Perl changed at release 5.18, and PCRE followed at
1345 release 8.34. "Space" and \es now match the same set of characters.
1346 .P
1347 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
1348 5.8. Another Perl extension is negation, which is indicated by a ^ character
1349 after the colon. For example,
1350 .sp
1351 [12[:^digit:]]
1352 .sp
1353 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
1354 syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
1355 supported, and an error is given if they are encountered.
1356 .P
1357 By default, in UTF modes, characters with values greater than 128 do not match
1358 any of the POSIX character classes. However, if the PCRE_UCP option is passed
1359 to \fBpcre_compile()\fP, some of the classes are changed so that Unicode
1360 character properties are used. This is achieved by replacing certain POSIX
1361 classes by other sequences, as follows:
1362 .sp
1363 [:alnum:] becomes \ep{Xan}
1364 [:alpha:] becomes \ep{L}
1365 [:blank:] becomes \eh
1366 [:digit:] becomes \ep{Nd}
1367 [:lower:] becomes \ep{Ll}
1368 [:space:] becomes \ep{Xps}
1369 [:upper:] becomes \ep{Lu}
1370 [:word:] becomes \ep{Xwd}
1371 .sp
1372 Negated versions, such as [:^alpha:] use \eP instead of \ep. Three other POSIX
1373 classes are handled specially in UCP mode:
1374 .TP 10
1375 [:graph:]
1376 This matches characters that have glyphs that mark the page when printed. In
1377 Unicode property terms, it matches all characters with the L, M, N, P, S, or Cf
1378 properties, except for:
1379 .sp
1380 U+061C Arabic Letter Mark
1381 U+180E Mongolian Vowel Separator
1382 U+2066 - U+2069 Various "isolate"s
1383 .sp
1384 .TP 10
1385 [:print:]
1386 This matches the same characters as [:graph:] plus space characters that are
1387 not controls, that is, characters with the Zs property.
1388 .TP 10
1389 [:punct:]
1390 This matches all characters that have the Unicode P (punctuation) property,
1391 plus those characters whose code points are less than 128 that have the S
1392 (Symbol) property.
1393 .P
1394 The other POSIX classes are unchanged, and match only characters with code
1395 points less than 128.
1396 .
1397 .
1399 .rs
1400 .sp
1401 Vertical bar characters are used to separate alternative patterns. For example,
1402 the pattern
1403 .sp
1404 gilbert|sullivan
1405 .sp
1406 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
1407 and an empty alternative is permitted (matching the empty string). The matching
1408 process tries each alternative in turn, from left to right, and the first one
1409 that succeeds is used. If the alternatives are within a subpattern
1410 .\" HTML <a href="#subpattern">
1411 .\" </a>
1412 (defined below),
1413 .\"
1414 "succeeds" means matching the rest of the main pattern as well as the
1415 alternative in the subpattern.
1416 .
1417 .
1419 .rs
1420 .sp
1421 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1422 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1423 the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1424 The option letters are
1425 .sp
1426 i for PCRE_CASELESS
1428 s for PCRE_DOTALL
1429 x for PCRE_EXTENDED
1430 .sp
1431 For example, (?im) sets caseless, multiline matching. It is also possible to
1432 unset these options by preceding the letter with a hyphen, and a combined
1433 setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1434 PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1435 permitted. If a letter appears both before and after the hyphen, the option is
1436 unset.
1437 .P
1438 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1439 changed in the same way as the Perl-compatible options by using the characters
1440 J, U and X respectively.
1441 .P
1442 When one of these option changes occurs at top level (that is, not inside
1443 subpattern parentheses), the change applies to the remainder of the pattern
1444 that follows. If the change is placed right at the start of a pattern, PCRE
1445 extracts it into the global options (and it will therefore show up in data
1446 extracted by the \fBpcre_fullinfo()\fP function).
1447 .P
1448 An option change within a subpattern (see below for a description of
1449 subpatterns) affects only that part of the subpattern that follows it, so
1450 .sp
1451 (a(?i)b)c
1452 .sp
1453 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1454 By this means, options can be made to have different settings in different
1455 parts of the pattern. Any changes made in one alternative do carry on
1456 into subsequent branches within the same subpattern. For example,
1457 .sp
1458 (a(?i)b|c)
1459 .sp
1460 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1461 branch is abandoned before the option setting. This is because the effects of
1462 option settings happen at compile time. There would be some very weird
1463 behaviour otherwise.
1464 .P
1465 \fBNote:\fP There are other PCRE-specific options that can be set by the
1466 application when the compiling or matching functions are called. In some cases
1467 the pattern can contain special leading sequences such as (*CRLF) to override
1468 what the application has set or what has been defaulted. Details are given in
1469 the section entitled
1470 .\" HTML <a href="#newlineseq">
1471 .\" </a>
1472 "Newline sequences"
1473 .\"
1474 above. There are also the (*UTF8), (*UTF16),(*UTF32), and (*UCP) leading
1475 sequences that can be used to set UTF and Unicode property modes; they are
1476 equivalent to setting the PCRE_UTF8, PCRE_UTF16, PCRE_UTF32 and the PCRE_UCP
1477 options, respectively. The (*UTF) sequence is a generic version that can be
1478 used with any of the libraries. However, the application can set the
1479 PCRE_NEVER_UTF option, which locks out the use of the (*UTF) sequences.
1480 .
1481 .
1482 .\" HTML <a name="subpattern"></a>
1484 .rs
1485 .sp
1486 Subpatterns are delimited by parentheses (round brackets), which can be nested.
1487 Turning part of a pattern into a subpattern does two things:
1488 .sp
1489 1. It localizes a set of alternatives. For example, the pattern
1490 .sp
1491 cat(aract|erpillar|)
1492 .sp
1493 matches "cataract", "caterpillar", or "cat". Without the parentheses, it would
1494 match "cataract", "erpillar" or an empty string.
1495 .sp
1496 2. It sets up the subpattern as a capturing subpattern. This means that, when
1497 the whole pattern matches, that portion of the subject string that matched the
1498 subpattern is passed back to the caller via the \fIovector\fP argument of the
1499 matching function. (This applies only to the traditional matching functions;
1500 the DFA matching functions do not support capturing.)
1501 .P
1502 Opening parentheses are counted from left to right (starting from 1) to obtain
1503 numbers for the capturing subpatterns. For example, if the string "the red
1504 king" is matched against the pattern
1505 .sp
1506 the ((red|white) (king|queen))
1507 .sp
1508 the captured substrings are "red king", "red", and "king", and are numbered 1,
1509 2, and 3, respectively.
1510 .P
1511 The fact that plain parentheses fulfil two functions is not always helpful.
1512 There are often times when a grouping subpattern is required without a
1513 capturing requirement. If an opening parenthesis is followed by a question mark
1514 and a colon, the subpattern does not do any capturing, and is not counted when
1515 computing the number of any subsequent capturing subpatterns. For example, if
1516 the string "the white queen" is matched against the pattern
1517 .sp
1518 the ((?:red|white) (king|queen))
1519 .sp
1520 the captured substrings are "white queen" and "queen", and are numbered 1 and
1521 2. The maximum number of capturing subpatterns is 65535.
1522 .P
1523 As a convenient shorthand, if any option settings are required at the start of
1524 a non-capturing subpattern, the option letters may appear between the "?" and
1525 the ":". Thus the two patterns
1526 .sp
1527 (?i:saturday|sunday)
1528 (?:(?i)saturday|sunday)
1529 .sp
1530 match exactly the same set of strings. Because alternative branches are tried
1531 from left to right, and options are not reset until the end of the subpattern
1532 is reached, an option setting in one branch does affect subsequent branches, so
1533 the above patterns match "SUNDAY" as well as "Saturday".
1534 .
1535 .
1536 .\" HTML <a name="dupsubpatternnumber"></a>
1538 .rs
1539 .sp
1540 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1541 the same numbers for its capturing parentheses. Such a subpattern starts with
1542 (?| and is itself a non-capturing subpattern. For example, consider this
1543 pattern:
1544 .sp
1545 (?|(Sat)ur|(Sun))day
1546 .sp
1547 Because the two alternatives are inside a (?| group, both sets of capturing
1548 parentheses are numbered one. Thus, when the pattern matches, you can look
1549 at captured substring number one, whichever alternative matched. This construct
1550 is useful when you want to capture part, but not all, of one of a number of
1551 alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1552 number is reset at the start of each branch. The numbers of any capturing
1553 parentheses that follow the subpattern start after the highest number used in
1554 any branch. The following example is taken from the Perl documentation. The
1555 numbers underneath show in which buffer the captured content will be stored.
1556 .sp
1557 # before ---------------branch-reset----------- after
1558 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1559 # 1 2 2 3 2 3 4
1560 .sp
1561 A back reference to a numbered subpattern uses the most recent value that is
1562 set for that number by any subpattern. The following pattern matches "abcabc"
1563 or "defdef":
1564 .sp
1565 /(?|(abc)|(def))\e1/
1566 .sp
1567 In contrast, a subroutine call to a numbered subpattern always refers to the
1568 first one in the pattern with the given number. The following pattern matches
1569 "abcabc" or "defabc":
1570 .sp
1571 /(?|(abc)|(def))(?1)/
1572 .sp
1573 If a
1574 .\" HTML <a href="#conditions">
1575 .\" </a>
1576 condition test
1577 .\"
1578 for a subpattern's having matched refers to a non-unique number, the test is
1579 true if any of the subpatterns of that number have matched.
1580 .P
1581 An alternative approach to using this "branch reset" feature is to use
1582 duplicate named subpatterns, as described in the next section.
1583 .
1584 .
1586 .rs
1587 .sp
1588 Identifying capturing parentheses by number is simple, but it can be very hard
1589 to keep track of the numbers in complicated regular expressions. Furthermore,
1590 if an expression is modified, the numbers may change. To help with this
1591 difficulty, PCRE supports the naming of subpatterns. This feature was not
1592 added to Perl until release 5.10. Python had the feature earlier, and PCRE
1593 introduced it at release 4.0, using the Python syntax. PCRE now supports both
1594 the Perl and the Python syntax. Perl allows identically numbered subpatterns to
1595 have different names, but PCRE does not.
1596 .P
1597 In PCRE, a subpattern can be named in one of three ways: (?<name>...) or
1598 (?'name'...) as in Perl, or (?P<name>...) as in Python. References to capturing
1599 parentheses from other parts of the pattern, such as
1600 .\" HTML <a href="#backreferences">
1601 .\" </a>
1602 back references,
1603 .\"
1604 .\" HTML <a href="#recursion">
1605 .\" </a>
1606 recursion,
1607 .\"
1608 and
1609 .\" HTML <a href="#conditions">
1610 .\" </a>
1611 conditions,
1612 .\"
1613 can be made by name as well as by number.
1614 .P
1615 Names consist of up to 32 alphanumeric characters and underscores, but must
1616 start with a non-digit. Named capturing parentheses are still allocated numbers
1617 as well as names, exactly as if the names were not present. The PCRE API
1618 provides function calls for extracting the name-to-number translation table
1619 from a compiled pattern. There is also a convenience function for extracting a
1620 captured substring by name.
1621 .P
1622 By default, a name must be unique within a pattern, but it is possible to relax
1623 this constraint by setting the PCRE_DUPNAMES option at compile time. (Duplicate
1624 names are also always permitted for subpatterns with the same number, set up as
1625 described in the previous section.) Duplicate names can be useful for patterns
1626 where only one instance of the named parentheses can match. Suppose you want to
1627 match the name of a weekday, either as a 3-letter abbreviation or as the full
1628 name, and in both cases you want to extract the abbreviation. This pattern
1629 (ignoring the line breaks) does the job:
1630 .sp
1631 (?<DN>Mon|Fri|Sun)(?:day)?|
1632 (?<DN>Tue)(?:sday)?|
1633 (?<DN>Wed)(?:nesday)?|
1634 (?<DN>Thu)(?:rsday)?|
1635 (?<DN>Sat)(?:urday)?
1636 .sp
1637 There are five capturing substrings, but only one is ever set after a match.
1638 (An alternative way of solving this problem is to use a "branch reset"
1639 subpattern, as described in the previous section.)
1640 .P
1641 The convenience function for extracting the data by name returns the substring
1642 for the first (and in this example, the only) subpattern of that name that
1643 matched. This saves searching to find which numbered subpattern it was.
1644 .P
1645 If you make a back reference to a non-unique named subpattern from elsewhere in
1646 the pattern, the subpatterns to which the name refers are checked in the order
1647 in which they appear in the overall pattern. The first one that is set is used
1648 for the reference. For example, this pattern matches both "foofoo" and
1649 "barbar" but not "foobar" or "barfoo":
1650 .sp
1651 (?:(?<n>foo)|(?<n>bar))\k<n>
1652 .sp
1653 .P
1654 If you make a subroutine call to a non-unique named subpattern, the one that
1655 corresponds to the first occurrence of the name is used. In the absence of
1656 duplicate numbers (see the previous section) this is the one with the lowest
1657 number.
1658 .P
1659 If you use a named reference in a condition
1660 test (see the
1661 .\"
1662 .\" HTML <a href="#conditions">
1663 .\" </a>
1664 section about conditions
1665 .\"
1666 below), either to check whether a subpattern has matched, or to check for
1667 recursion, all subpatterns with the same name are tested. If the condition is
1668 true for any one of them, the overall condition is true. This is the same
1669 behaviour as testing by number. For further details of the interfaces for
1670 handling named subpatterns, see the
1671 .\" HREF
1672 \fBpcreapi\fP
1673 .\"
1674 documentation.
1675 .P
1676 \fBWarning:\fP You cannot use different names to distinguish between two
1677 subpatterns with the same number because PCRE uses only the numbers when
1678 matching. For this reason, an error is given at compile time if different names
1679 are given to subpatterns with the same number. However, you can always give the
1680 same name to subpatterns with the same number, even when PCRE_DUPNAMES is not
1681 set.
1682 .
1683 .
1685 .rs
1686 .sp
1687 Repetition is specified by quantifiers, which can follow any of the following
1688 items:
1689 .sp
1690 a literal data character
1691 the dot metacharacter
1692 the \eC escape sequence
1693 the \eX escape sequence
1694 the \eR escape sequence
1695 an escape such as \ed or \epL that matches a single character
1696 a character class
1697 a back reference (see next section)
1698 a parenthesized subpattern (including assertions)
1699 a subroutine call to a subpattern (recursive or otherwise)
1700 .sp
1701 The general repetition quantifier specifies a minimum and maximum number of
1702 permitted matches, by giving the two numbers in curly brackets (braces),
1703 separated by a comma. The numbers must be less than 65536, and the first must
1704 be less than or equal to the second. For example:
1705 .sp
1706 z{2,4}
1707 .sp
1708 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1709 character. If the second number is omitted, but the comma is present, there is
1710 no upper limit; if the second number and the comma are both omitted, the
1711 quantifier specifies an exact number of required matches. Thus
1712 .sp
1713 [aeiou]{3,}
1714 .sp
1715 matches at least 3 successive vowels, but may match many more, while
1716 .sp
1717 \ed{8}
1718 .sp
1719 matches exactly 8 digits. An opening curly bracket that appears in a position
1720 where a quantifier is not allowed, or one that does not match the syntax of a
1721 quantifier, is taken as a literal character. For example, {,6} is not a
1722 quantifier, but a literal string of four characters.
1723 .P
1724 In UTF modes, quantifiers apply to characters rather than to individual data
1725 units. Thus, for example, \ex{100}{2} matches two characters, each of
1726 which is represented by a two-byte sequence in a UTF-8 string. Similarly,
1727 \eX{3} matches three Unicode extended grapheme clusters, each of which may be
1728 several data units long (and they may be of different lengths).
1729 .P
1730 The quantifier {0} is permitted, causing the expression to behave as if the
1731 previous item and the quantifier were not present. This may be useful for
1732 subpatterns that are referenced as
1733 .\" HTML <a href="#subpatternsassubroutines">
1734 .\" </a>
1735 subroutines
1736 .\"
1737 from elsewhere in the pattern (but see also the section entitled
1738 .\" HTML <a href="#subdefine">
1739 .\" </a>
1740 "Defining subpatterns for use by reference only"
1741 .\"
1742 below). Items other than subpatterns that have a {0} quantifier are omitted
1743 from the compiled pattern.
1744 .P
1745 For convenience, the three most common quantifiers have single-character
1746 abbreviations:
1747 .sp
1748 * is equivalent to {0,}
1749 + is equivalent to {1,}
1750 ? is equivalent to {0,1}
1751 .sp
1752 It is possible to construct infinite loops by following a subpattern that can
1753 match no characters with a quantifier that has no upper limit, for example:
1754 .sp
1755 (a?)*
1756 .sp
1757 Earlier versions of Perl and PCRE used to give an error at compile time for
1758 such patterns. However, because there are cases where this can be useful, such
1759 patterns are now accepted, but if any repetition of the subpattern does in fact
1760 match no characters, the loop is forcibly broken.
1761 .P
1762 By default, the quantifiers are "greedy", that is, they match as much as
1763 possible (up to the maximum number of permitted times), without causing the
1764 rest of the pattern to fail. The classic example of where this gives problems
1765 is in trying to match comments in C programs. These appear between /* and */
1766 and within the comment, individual * and / characters may appear. An attempt to
1767 match C comments by applying the pattern
1768 .sp
1769 /\e*.*\e*/
1770 .sp
1771 to the string
1772 .sp
1773 /* first comment */ not comment /* second comment */
1774 .sp
1775 fails, because it matches the entire string owing to the greediness of the .*
1776 item.
1777 .P
1778 However, if a quantifier is followed by a question mark, it ceases to be
1779 greedy, and instead matches the minimum number of times possible, so the
1780 pattern
1781 .sp
1782 /\e*.*?\e*/
1783 .sp
1784 does the right thing with the C comments. The meaning of the various
1785 quantifiers is not otherwise changed, just the preferred number of matches.
1786 Do not confuse this use of question mark with its use as a quantifier in its
1787 own right. Because it has two uses, it can sometimes appear doubled, as in
1788 .sp
1789 \ed??\ed
1790 .sp
1791 which matches one digit by preference, but can match two if that is the only
1792 way the rest of the pattern matches.
1793 .P
1794 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1795 the quantifiers are not greedy by default, but individual ones can be made
1796 greedy by following them with a question mark. In other words, it inverts the
1797 default behaviour.
1798 .P
1799 When a parenthesized subpattern is quantified with a minimum repeat count that
1800 is greater than 1 or with a limited maximum, more memory is required for the
1801 compiled pattern, in proportion to the size of the minimum or maximum.
1802 .P
1803 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1804 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1805 implicitly anchored, because whatever follows will be tried against every
1806 character position in the subject string, so there is no point in retrying the
1807 overall match at any position after the first. PCRE normally treats such a
1808 pattern as though it were preceded by \eA.
1809 .P
1810 In cases where it is known that the subject string contains no newlines, it is
1811 worth setting PCRE_DOTALL in order to obtain this optimization, or
1812 alternatively using ^ to indicate anchoring explicitly.
1813 .P
1814 However, there are some cases where the optimization cannot be used. When .*
1815 is inside capturing parentheses that are the subject of a back reference
1816 elsewhere in the pattern, a match at the start may fail where a later one
1817 succeeds. Consider, for example:
1818 .sp
1819 (.*)abc\e1
1820 .sp
1821 If the subject is "xyz123abc123" the match point is the fourth character. For
1822 this reason, such a pattern is not implicitly anchored.
1823 .P
1824 Another case where implicit anchoring is not applied is when the leading .* is
1825 inside an atomic group. Once again, a match at the start may fail where a later
1826 one succeeds. Consider this pattern:
1827 .sp
1828 (?>.*?a)b
1829 .sp
1830 It matches "ab" in the subject "aab". The use of the backtracking control verbs
1831 (*PRUNE) and (*SKIP) also disable this optimization.
1832 .P
1833 When a capturing subpattern is repeated, the value captured is the substring
1834 that matched the final iteration. For example, after
1835 .sp
1836 (tweedle[dume]{3}\es*)+
1837 .sp
1838 has matched "tweedledum tweedledee" the value of the captured substring is
1839 "tweedledee". However, if there are nested capturing subpatterns, the
1840 corresponding captured values may have been set in previous iterations. For
1841 example, after
1842 .sp
1843 /(a|(b))+/
1844 .sp
1845 matches "aba" the value of the second captured substring is "b".
1846 .
1847 .
1848 .\" HTML <a name="atomicgroup"></a>
1850 .rs
1851 .sp
1852 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1853 repetition, failure of what follows normally causes the repeated item to be
1854 re-evaluated to see if a different number of repeats allows the rest of the
1855 pattern to match. Sometimes it is useful to prevent this, either to change the
1856 nature of the match, or to cause it fail earlier than it otherwise might, when
1857 the author of the pattern knows there is no point in carrying on.
1858 .P
1859 Consider, for example, the pattern \ed+foo when applied to the subject line
1860 .sp
1861 123456bar
1862 .sp
1863 After matching all 6 digits and then failing to match "foo", the normal
1864 action of the matcher is to try again with only 5 digits matching the \ed+
1865 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1866 (a term taken from Jeffrey Friedl's book) provides the means for specifying
1867 that once a subpattern has matched, it is not to be re-evaluated in this way.
1868 .P
1869 If we use atomic grouping for the previous example, the matcher gives up
1870 immediately on failing to match "foo" the first time. The notation is a kind of
1871 special parenthesis, starting with (?> as in this example:
1872 .sp
1873 (?>\ed+)foo
1874 .sp
1875 This kind of parenthesis "locks up" the part of the pattern it contains once
1876 it has matched, and a failure further into the pattern is prevented from
1877 backtracking into it. Backtracking past it to previous items, however, works as
1878 normal.
1879 .P
1880 An alternative description is that a subpattern of this type matches the string
1881 of characters that an identical standalone pattern would match, if anchored at
1882 the current point in the subject string.
1883 .P
1884 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1885 the above example can be thought of as a maximizing repeat that must swallow
1886 everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1887 number of digits they match in order to make the rest of the pattern match,
1888 (?>\ed+) can only match an entire sequence of digits.
1889 .P
1890 Atomic groups in general can of course contain arbitrarily complicated
1891 subpatterns, and can be nested. However, when the subpattern for an atomic
1892 group is just a single repeated item, as in the example above, a simpler
1893 notation, called a "possessive quantifier" can be used. This consists of an
1894 additional + character following a quantifier. Using this notation, the
1895 previous example can be rewritten as
1896 .sp
1897 \ed++foo
1898 .sp
1899 Note that a possessive quantifier can be used with an entire group, for
1900 example:
1901 .sp
1902 (abc|xyz){2,3}+
1903 .sp
1904 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1905 option is ignored. They are a convenient notation for the simpler forms of
1906 atomic group. However, there is no difference in the meaning of a possessive
1907 quantifier and the equivalent atomic group, though there may be a performance
1908 difference; possessive quantifiers should be slightly faster.
1909 .P
1910 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1911 Jeffrey Friedl originated the idea (and the name) in the first edition of his
1912 book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1913 package, and PCRE copied it from there. It ultimately found its way into Perl
1914 at release 5.10.
1915 .P
1916 PCRE has an optimization that automatically "possessifies" certain simple
1917 pattern constructs. For example, the sequence A+B is treated as A++B because
1918 there is no point in backtracking into a sequence of A's when B must follow.
1919 .P
1920 When a pattern contains an unlimited repeat inside a subpattern that can itself
1921 be repeated an unlimited number of times, the use of an atomic group is the
1922 only way to avoid some failing matches taking a very long time indeed. The
1923 pattern
1924 .sp
1925 (\eD+|<\ed+>)*[!?]
1926 .sp
1927 matches an unlimited number of substrings that either consist of non-digits, or
1928 digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1929 quickly. However, if it is applied to
1930 .sp
1931 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1932 .sp
1933 it takes a long time before reporting failure. This is because the string can
1934 be divided between the internal \eD+ repeat and the external * repeat in a
1935 large number of ways, and all have to be tried. (The example uses [!?] rather
1936 than a single character at the end, because both PCRE and Perl have an
1937 optimization that allows for fast failure when a single character is used. They
1938 remember the last single character that is required for a match, and fail early
1939 if it is not present in the string.) If the pattern is changed so that it uses
1940 an atomic group, like this:
1941 .sp
1942 ((?>\eD+)|<\ed+>)*[!?]
1943 .sp
1944 sequences of non-digits cannot be broken, and failure happens quickly.
1945 .
1946 .
1947 .\" HTML <a name="backreferences"></a>
1949 .rs
1950 .sp
1951 Outside a character class, a backslash followed by a digit greater than 0 (and
1952 possibly further digits) is a back reference to a capturing subpattern earlier
1953 (that is, to its left) in the pattern, provided there have been that many
1954 previous capturing left parentheses.
1955 .P
1956 However, if the decimal number following the backslash is less than 10, it is
1957 always taken as a back reference, and causes an error only if there are not
1958 that many capturing left parentheses in the entire pattern. In other words, the
1959 parentheses that are referenced need not be to the left of the reference for
1960 numbers less than 10. A "forward back reference" of this type can make sense
1961 when a repetition is involved and the subpattern to the right has participated
1962 in an earlier iteration.
1963 .P
1964 It is not possible to have a numerical "forward back reference" to a subpattern
1965 whose number is 10 or more using this syntax because a sequence such as \e50 is
1966 interpreted as a character defined in octal. See the subsection entitled
1967 "Non-printing characters"
1968 .\" HTML <a href="#digitsafterbackslash">
1969 .\" </a>
1970 above
1971 .\"
1972 for further details of the handling of digits following a backslash. There is
1973 no such problem when named parentheses are used. A back reference to any
1974 subpattern is possible using named parentheses (see below).
1975 .P
1976 Another way of avoiding the ambiguity inherent in the use of digits following a
1977 backslash is to use the \eg escape sequence. This escape must be followed by an
1978 unsigned number or a negative number, optionally enclosed in braces. These
1979 examples are all identical:
1980 .sp
1981 (ring), \e1
1982 (ring), \eg1
1983 (ring), \eg{1}
1984 .sp
1985 An unsigned number specifies an absolute reference without the ambiguity that
1986 is present in the older syntax. It is also useful when literal digits follow
1987 the reference. A negative number is a relative reference. Consider this
1988 example:
1989 .sp
1990 (abc(def)ghi)\eg{-1}
1991 .sp
1992 The sequence \eg{-1} is a reference to the most recently started capturing
1993 subpattern before \eg, that is, is it equivalent to \e2 in this example.
1994 Similarly, \eg{-2} would be equivalent to \e1. The use of relative references
1995 can be helpful in long patterns, and also in patterns that are created by
1996 joining together fragments that contain references within themselves.
1997 .P
1998 A back reference matches whatever actually matched the capturing subpattern in
1999 the current subject string, rather than anything matching the subpattern
2000 itself (see
2001 .\" HTML <a href="#subpatternsassubroutines">
2002 .\" </a>
2003 "Subpatterns as subroutines"
2004 .\"
2005 below for a way of doing that). So the pattern
2006 .sp
2007 (sens|respons)e and \e1ibility
2008 .sp
2009 matches "sense and sensibility" and "response and responsibility", but not
2010 "sense and responsibility". If caseful matching is in force at the time of the
2011 back reference, the case of letters is relevant. For example,
2012 .sp
2013 ((?i)rah)\es+\e1
2014 .sp
2015 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
2016 capturing subpattern is matched caselessly.
2017 .P
2018 There are several different ways of writing back references to named
2019 subpatterns. The .NET syntax \ek{name} and the Perl syntax \ek<name> or
2020 \ek'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
2021 back reference syntax, in which \eg can be used for both numeric and named
2022 references, is also supported. We could rewrite the above example in any of
2023 the following ways:
2024 .sp
2025 (?<p1>(?i)rah)\es+\ek<p1>
2026 (?'p1'(?i)rah)\es+\ek{p1}
2027 (?P<p1>(?i)rah)\es+(?P=p1)
2028 (?<p1>(?i)rah)\es+\eg{p1}
2029 .sp
2030 A subpattern that is referenced by name may appear in the pattern before or
2031 after the reference.
2032 .P
2033 There may be more than one back reference to the same subpattern. If a
2034 subpattern has not actually been used in a particular match, any back
2035 references to it always fail by default. For example, the pattern
2036 .sp
2037 (a|(bc))\e2
2038 .sp
2039 always fails if it starts to match "a" rather than "bc". However, if the
2040 PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back reference to an
2041 unset value matches an empty string.
2042 .P
2043 Because there may be many capturing parentheses in a pattern, all digits
2044 following a backslash are taken as part of a potential back reference number.
2045 If the pattern continues with a digit character, some delimiter must be used to
2046 terminate the back reference. If the PCRE_EXTENDED option is set, this can be
2047 white space. Otherwise, the \eg{ syntax or an empty comment (see
2048 .\" HTML <a href="#comments">
2049 .\" </a>
2050 "Comments"
2051 .\"
2052 below) can be used.
2053 .
2054 .SS "Recursive back references"
2055 .rs
2056 .sp
2057 A back reference that occurs inside the parentheses to which it refers fails
2058 when the subpattern is first used, so, for example, (a\e1) never matches.
2059 However, such references can be useful inside repeated subpatterns. For
2060 example, the pattern
2061 .sp
2062 (a|b\e1)+
2063 .sp
2064 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
2065 the subpattern, the back reference matches the character string corresponding
2066 to the previous iteration. In order for this to work, the pattern must be such
2067 that the first iteration does not need to match the back reference. This can be
2068 done using alternation, as in the example above, or by a quantifier with a
2069 minimum of zero.
2070 .P
2071 Back references of this type cause the group that they reference to be treated
2072 as an
2073 .\" HTML <a href="#atomicgroup">
2074 .\" </a>
2075 atomic group.
2076 .\"
2077 Once the whole group has been matched, a subsequent matching failure cannot
2078 cause backtracking into the middle of the group.
2079 .
2080 .
2081 .\" HTML <a name="bigassertions"></a>
2083 .rs
2084 .sp
2085 An assertion is a test on the characters following or preceding the current
2086 matching point that does not actually consume any characters. The simple
2087 assertions coded as \eb, \eB, \eA, \eG, \eZ, \ez, ^ and $ are described
2088 .\" HTML <a href="#smallassertions">
2089 .\" </a>
2090 above.
2091 .\"
2092 .P
2093 More complicated assertions are coded as subpatterns. There are two kinds:
2094 those that look ahead of the current position in the subject string, and those
2095 that look behind it. An assertion subpattern is matched in the normal way,
2096 except that it does not cause the current matching position to be changed.
2097 .P
2098 Assertion subpatterns are not capturing subpatterns. If such an assertion
2099 contains capturing subpatterns within it, these are counted for the purposes of
2100 numbering the capturing subpatterns in the whole pattern. However, substring
2101 capturing is carried out only for positive assertions. (Perl sometimes, but not
2102 always, does do capturing in negative assertions.)
2103 .P
2104 For compatibility with Perl, assertion subpatterns may be repeated; though
2105 it makes no sense to assert the same thing several times, the side effect of
2106 capturing parentheses may occasionally be useful. In practice, there only three
2107 cases:
2108 .sp
2109 (1) If the quantifier is {0}, the assertion is never obeyed during matching.
2110 However, it may contain internal capturing parenthesized groups that are called
2111 from elsewhere via the
2112 .\" HTML <a href="#subpatternsassubroutines">
2113 .\" </a>
2114 subroutine mechanism.
2115 .\"
2116 .sp
2117 (2) If quantifier is {0,n} where n is greater than zero, it is treated as if it
2118 were {0,1}. At run time, the rest of the pattern match is tried with and
2119 without the assertion, the order depending on the greediness of the quantifier.
2120 .sp
2121 (3) If the minimum repetition is greater than zero, the quantifier is ignored.
2122 The assertion is obeyed just once when encountered during matching.
2123 .
2124 .
2125 .SS "Lookahead assertions"
2126 .rs
2127 .sp
2128 Lookahead assertions start with (?= for positive assertions and (?! for
2129 negative assertions. For example,
2130 .sp
2131 \ew+(?=;)
2132 .sp
2133 matches a word followed by a semicolon, but does not include the semicolon in
2134 the match, and
2135 .sp
2136 foo(?!bar)
2137 .sp
2138 matches any occurrence of "foo" that is not followed by "bar". Note that the
2139 apparently similar pattern
2140 .sp
2141 (?!foo)bar
2142 .sp
2143 does not find an occurrence of "bar" that is preceded by something other than
2144 "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
2145 (?!foo) is always true when the next three characters are "bar". A
2146 lookbehind assertion is needed to achieve the other effect.
2147 .P
2148 If you want to force a matching failure at some point in a pattern, the most
2149 convenient way to do it is with (?!) because an empty string always matches, so
2150 an assertion that requires there not to be an empty string must always fail.
2151 The backtracking control verb (*FAIL) or (*F) is a synonym for (?!).
2152 .
2153 .
2154 .\" HTML <a name="lookbehind"></a>
2155 .SS "Lookbehind assertions"
2156 .rs
2157 .sp
2158 Lookbehind assertions start with (?<= for positive assertions and (?<! for
2159 negative assertions. For example,
2160 .sp
2161 (?<!foo)bar
2162 .sp
2163 does find an occurrence of "bar" that is not preceded by "foo". The contents of
2164 a lookbehind assertion are restricted such that all the strings it matches must
2165 have a fixed length. However, if there are several top-level alternatives, they
2166 do not all have to have the same fixed length. Thus
2167 .sp
2168 (?<=bullock|donkey)
2169 .sp
2170 is permitted, but
2171 .sp
2172 (?<!dogs?|cats?)
2173 .sp
2174 causes an error at compile time. Branches that match different length strings
2175 are permitted only at the top level of a lookbehind assertion. This is an
2176 extension compared with Perl, which requires all branches to match the same
2177 length of string. An assertion such as
2178 .sp
2179 (?<=ab(c|de))
2180 .sp
2181 is not permitted, because its single top-level branch can match two different
2182 lengths, but it is acceptable to PCRE if rewritten to use two top-level
2183 branches:
2184 .sp
2185 (?<=abc|abde)
2186 .sp
2187 In some cases, the escape sequence \eK
2188 .\" HTML <a href="#resetmatchstart">
2189 .\" </a>
2190 (see above)
2191 .\"
2192 can be used instead of a lookbehind assertion to get round the fixed-length
2193 restriction.
2194 .P
2195 The implementation of lookbehind assertions is, for each alternative, to
2196 temporarily move the current position back by the fixed length and then try to
2197 match. If there are insufficient characters before the current position, the
2198 assertion fails.
2199 .P
2200 In a UTF mode, PCRE does not allow the \eC escape (which matches a single data
2201 unit even in a UTF mode) to appear in lookbehind assertions, because it makes
2202 it impossible to calculate the length of the lookbehind. The \eX and \eR
2203 escapes, which can match different numbers of data units, are also not
2204 permitted.
2205 .P
2206 .\" HTML <a href="#subpatternsassubroutines">
2207 .\" </a>
2208 "Subroutine"
2209 .\"
2210 calls (see below) such as (?2) or (?&X) are permitted in lookbehinds, as long
2211 as the subpattern matches a fixed-length string.
2212 .\" HTML <a href="#recursion">
2213 .\" </a>
2214 Recursion,
2215 .\"
2216 however, is not supported.
2217 .P
2218 Possessive quantifiers can be used in conjunction with lookbehind assertions to
2219 specify efficient matching of fixed-length strings at the end of subject
2220 strings. Consider a simple pattern such as
2221 .sp
2222 abcd$
2223 .sp
2224 when applied to a long string that does not match. Because matching proceeds
2225 from left to right, PCRE will look for each "a" in the subject and then see if
2226 what follows matches the rest of the pattern. If the pattern is specified as
2227 .sp
2228 ^.*abcd$
2229 .sp
2230 the initial .* matches the entire string at first, but when this fails (because
2231 there is no following "a"), it backtracks to match all but the last character,
2232 then all but the last two characters, and so on. Once again the search for "a"
2233 covers the entire string, from right to left, so we are no better off. However,
2234 if the pattern is written as
2235 .sp
2236 ^.*+(?<=abcd)
2237 .sp
2238 there can be no backtracking for the .*+ item; it can match only the entire
2239 string. The subsequent lookbehind assertion does a single test on the last four
2240 characters. If it fails, the match fails immediately. For long strings, this
2241 approach makes a significant difference to the processing time.
2242 .
2243 .
2244 .SS "Using multiple assertions"
2245 .rs
2246 .sp
2247 Several assertions (of any sort) may occur in succession. For example,
2248 .sp
2249 (?<=\ed{3})(?<!999)foo
2250 .sp
2251 matches "foo" preceded by three digits that are not "999". Notice that each of
2252 the assertions is applied independently at the same point in the subject
2253 string. First there is a check that the previous three characters are all
2254 digits, and then there is a check that the same three characters are not "999".
2255 This pattern does \fInot\fP match "foo" preceded by six characters, the first
2256 of which are digits and the last three of which are not "999". For example, it
2257 doesn't match "123abcfoo". A pattern to do that is
2258 .sp
2259 (?<=\ed{3}...)(?<!999)foo
2260 .sp
2261 This time the first assertion looks at the preceding six characters, checking
2262 that the first three are digits, and then the second assertion checks that the
2263 preceding three characters are not "999".
2264 .P
2265 Assertions can be nested in any combination. For example,
2266 .sp
2267 (?<=(?<!foo)bar)baz
2268 .sp
2269 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
2270 preceded by "foo", while
2271 .sp
2272 (?<=\ed{3}(?!999)...)foo
2273 .sp
2274 is another pattern that matches "foo" preceded by three digits and any three
2275 characters that are not "999".
2276 .
2277 .
2278 .\" HTML <a name="conditions"></a>
2280 .rs
2281 .sp
2282 It is possible to cause the matching process to obey a subpattern
2283 conditionally or to choose between two alternative subpatterns, depending on
2284 the result of an assertion, or whether a specific capturing subpattern has
2285 already been matched. The two possible forms of conditional subpattern are:
2286 .sp
2287 (?(condition)yes-pattern)
2288 (?(condition)yes-pattern|no-pattern)
2289 .sp
2290 If the condition is satisfied, the yes-pattern is used; otherwise the
2291 no-pattern (if present) is used. If there are more than two alternatives in the
2292 subpattern, a compile-time error occurs. Each of the two alternatives may
2293 itself contain nested subpatterns of any form, including conditional
2294 subpatterns; the restriction to two alternatives applies only at the level of
2295 the condition. This pattern fragment is an example where the alternatives are
2296 complex:
2297 .sp
2298 (?(1) (A|B|C) | (D | (?(2)E|F) | E) )
2299 .sp
2300 .P
2301 There are four kinds of condition: references to subpatterns, references to
2302 recursion, a pseudo-condition called DEFINE, and assertions.
2303 .
2304 .SS "Checking for a used subpattern by number"
2305 .rs
2306 .sp
2307 If the text between the parentheses consists of a sequence of digits, the
2308 condition is true if a capturing subpattern of that number has previously
2309 matched. If there is more than one capturing subpattern with the same number
2310 (see the earlier
2311 .\"
2312 .\" HTML <a href="#recursion">
2313 .\" </a>
2314 section about duplicate subpattern numbers),
2315 .\"
2316 the condition is true if any of them have matched. An alternative notation is
2317 to precede the digits with a plus or minus sign. In this case, the subpattern
2318 number is relative rather than absolute. The most recently opened parentheses
2319 can be referenced by (?(-1), the next most recent by (?(-2), and so on. Inside
2320 loops it can also make sense to refer to subsequent groups. The next
2321 parentheses to be opened can be referenced as (?(+1), and so on. (The value
2322 zero in any of these forms is not used; it provokes a compile-time error.)
2323 .P
2324 Consider the following pattern, which contains non-significant white space to
2325 make it more readable (assume the PCRE_EXTENDED option) and to divide it into
2326 three parts for ease of discussion:
2327 .sp
2328 ( \e( )? [^()]+ (?(1) \e) )
2329 .sp
2330 The first part matches an optional opening parenthesis, and if that
2331 character is present, sets it as the first captured substring. The second part
2332 matches one or more characters that are not parentheses. The third part is a
2333 conditional subpattern that tests whether or not the first set of parentheses
2334 matched. If they did, that is, if subject started with an opening parenthesis,
2335 the condition is true, and so the yes-pattern is executed and a closing
2336 parenthesis is required. Otherwise, since no-pattern is not present, the
2337 subpattern matches nothing. In other words, this pattern matches a sequence of
2338 non-parentheses, optionally enclosed in parentheses.
2339 .P
2340 If you were embedding this pattern in a larger one, you could use a relative
2341 reference:
2342 .sp
2343 ...other stuff... ( \e( )? [^()]+ (?(-1) \e) ) ...
2344 .sp
2345 This makes the fragment independent of the parentheses in the larger pattern.
2346 .
2347 .SS "Checking for a used subpattern by name"
2348 .rs
2349 .sp
2350 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a used
2351 subpattern by name. For compatibility with earlier versions of PCRE, which had
2352 this facility before Perl, the syntax (?(name)...) is also recognized.
2353 .P
2354 Rewriting the above example to use a named subpattern gives this:
2355 .sp
2356 (?<OPEN> \e( )? [^()]+ (?(<OPEN>) \e) )
2357 .sp
2358 If the name used in a condition of this kind is a duplicate, the test is
2359 applied to all subpatterns of the same name, and is true if any one of them has
2360 matched.
2361 .
2362 .SS "Checking for pattern recursion"
2363 .rs
2364 .sp
2365 If the condition is the string (R), and there is no subpattern with the name R,
2366 the condition is true if a recursive call to the whole pattern or any
2367 subpattern has been made. If digits or a name preceded by ampersand follow the
2368 letter R, for example:
2369 .sp
2370 (?(R3)...) or (?(R&name)...)
2371 .sp
2372 the condition is true if the most recent recursion is into a subpattern whose
2373 number or name is given. This condition does not check the entire recursion
2374 stack. If the name used in a condition of this kind is a duplicate, the test is
2375 applied to all subpatterns of the same name, and is true if any one of them is
2376 the most recent recursion.
2377 .P
2378 At "top level", all these recursion test conditions are false.
2379 .\" HTML <a href="#recursion">
2380 .\" </a>
2381 The syntax for recursive patterns
2382 .\"
2383 is described below.
2384 .
2385 .\" HTML <a name="subdefine"></a>
2386 .SS "Defining subpatterns for use by reference only"
2387 .rs
2388 .sp
2389 If the condition is the string (DEFINE), and there is no subpattern with the
2390 name DEFINE, the condition is always false. In this case, there may be only one
2391 alternative in the subpattern. It is always skipped if control reaches this
2392 point in the pattern; the idea of DEFINE is that it can be used to define
2393 subroutines that can be referenced from elsewhere. (The use of
2394 .\" HTML <a href="#subpatternsassubroutines">
2395 .\" </a>
2396 subroutines
2397 .\"
2398 is described below.) For example, a pattern to match an IPv4 address such as
2399 "" could be written like this (ignore white space and line
2400 breaks):
2401 .sp
2402 (?(DEFINE) (?<byte> 2[0-4]\ed | 25[0-5] | 1\ed\ed | [1-9]?\ed) )
2403 \eb (?&byte) (\e.(?&byte)){3} \eb
2404 .sp
2405 The first part of the pattern is a DEFINE group inside which a another group
2406 named "byte" is defined. This matches an individual component of an IPv4
2407 address (a number less than 256). When matching takes place, this part of the
2408 pattern is skipped because DEFINE acts like a false condition. The rest of the
2409 pattern uses references to the named group to match the four dot-separated
2410 components of an IPv4 address, insisting on a word boundary at each end.
2411 .
2412 .SS "Assertion conditions"
2413 .rs
2414 .sp
2415 If the condition is not in any of the above formats, it must be an assertion.
2416 This may be a positive or negative lookahead or lookbehind assertion. Consider
2417 this pattern, again containing non-significant white space, and with the two
2418 alternatives on the second line:
2419 .sp
2420 (?(?=[^a-z]*[a-z])
2421 \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
2422 .sp
2423 The condition is a positive lookahead assertion that matches an optional
2424 sequence of non-letters followed by a letter. In other words, it tests for the
2425 presence of at least one letter in the subject. If a letter is found, the
2426 subject is matched against the first alternative; otherwise it is matched
2427 against the second. This pattern matches strings in one of the two forms
2428 dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
2429 .
2430 .
2431 .\" HTML <a name="comments"></a>
2433 .rs
2434 .sp
2435 There are two ways of including comments in patterns that are processed by
2436 PCRE. In both cases, the start of the comment must not be in a character class,
2437 nor in the middle of any other sequence of related characters such as (?: or a
2438 subpattern name or number. The characters that make up a comment play no part
2439 in the pattern matching.
2440 .P
2441 The sequence (?# marks the start of a comment that continues up to the next
2442 closing parenthesis. Nested parentheses are not permitted. If the PCRE_EXTENDED
2443 option is set, an unescaped # character also introduces a comment, which in
2444 this case continues to immediately after the next newline character or
2445 character sequence in the pattern. Which characters are interpreted as newlines
2446 is controlled by the options passed to a compiling function or by a special
2447 sequence at the start of the pattern, as described in the section entitled
2448 .\" HTML <a href="#newlines">
2449 .\" </a>
2450 "Newline conventions"
2451 .\"
2452 above. Note that the end of this type of comment is a literal newline sequence
2453 in the pattern; escape sequences that happen to represent a newline do not
2454 count. For example, consider this pattern when PCRE_EXTENDED is set, and the
2455 default newline convention is in force:
2456 .sp
2457 abc #comment \en still comment
2458 .sp
2459 On encountering the # character, \fBpcre_compile()\fP skips along, looking for
2460 a newline in the pattern. The sequence \en is still literal at this stage, so
2461 it does not terminate the comment. Only an actual character with the code value
2462 0x0a (the default newline) does so.
2463 .
2464 .
2465 .\" HTML <a name="recursion"></a>
2467 .rs
2468 .sp
2469 Consider the problem of matching a string in parentheses, allowing for
2470 unlimited nested parentheses. Without the use of recursion, the best that can
2471 be done is to use a pattern that matches up to some fixed depth of nesting. It
2472 is not possible to handle an arbitrary nesting depth.
2473 .P
2474 For some time, Perl has provided a facility that allows regular expressions to
2475 recurse (amongst other things). It does this by interpolating Perl code in the
2476 expression at run time, and the code can refer to the expression itself. A Perl
2477 pattern using code interpolation to solve the parentheses problem can be
2478 created like this:
2479 .sp
2480 $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
2481 .sp
2482 The (?p{...}) item interpolates Perl code at run time, and in this case refers
2483 recursively to the pattern in which it appears.
2484 .P
2485 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
2486 supports special syntax for recursion of the entire pattern, and also for
2487 individual subpattern recursion. After its introduction in PCRE and Python,
2488 this kind of recursion was subsequently introduced into Perl at release 5.10.
2489 .P
2490 A special item that consists of (? followed by a number greater than zero and a
2491 closing parenthesis is a recursive subroutine call of the subpattern of the
2492 given number, provided that it occurs inside that subpattern. (If not, it is a
2493 .\" HTML <a href="#subpatternsassubroutines">
2494 .\" </a>
2495 non-recursive subroutine
2496 .\"
2497 call, which is described in the next section.) The special item (?R) or (?0) is
2498 a recursive call of the entire regular expression.
2499 .P
2500 This PCRE pattern solves the nested parentheses problem (assume the
2501 PCRE_EXTENDED option is set so that white space is ignored):
2502 .sp
2503 \e( ( [^()]++ | (?R) )* \e)
2504 .sp
2505 First it matches an opening parenthesis. Then it matches any number of
2506 substrings which can either be a sequence of non-parentheses, or a recursive
2507 match of the pattern itself (that is, a correctly parenthesized substring).
2508 Finally there is a closing parenthesis. Note the use of a possessive quantifier
2509 to avoid backtracking into sequences of non-parentheses.
2510 .P
2511 If this were part of a larger pattern, you would not want to recurse the entire
2512 pattern, so instead you could use this:
2513 .sp
2514 ( \e( ( [^()]++ | (?1) )* \e) )
2515 .sp
2516 We have put the pattern into parentheses, and caused the recursion to refer to
2517 them instead of the whole pattern.
2518 .P
2519 In a larger pattern, keeping track of parenthesis numbers can be tricky. This
2520 is made easier by the use of relative references. Instead of (?1) in the
2521 pattern above you can write (?-2) to refer to the second most recently opened
2522 parentheses preceding the recursion. In other words, a negative number counts
2523 capturing parentheses leftwards from the point at which it is encountered.
2524 .P
2525 It is also possible to refer to subsequently opened parentheses, by writing
2526 references such as (?+2). However, these cannot be recursive because the
2527 reference is not inside the parentheses that are referenced. They are always
2528 .\" HTML <a href="#subpatternsassubroutines">
2529 .\" </a>
2530 non-recursive subroutine
2531 .\"
2532 calls, as described in the next section.
2533 .P
2534 An alternative approach is to use named parentheses instead. The Perl syntax
2535 for this is (?&name); PCRE's earlier syntax (?P>name) is also supported. We
2536 could rewrite the above example as follows:
2537 .sp
2538 (?<pn> \e( ( [^()]++ | (?&pn) )* \e) )
2539 .sp
2540 If there is more than one subpattern with the same name, the earliest one is
2541 used.
2542 .P
2543 This particular example pattern that we have been looking at contains nested
2544 unlimited repeats, and so the use of a possessive quantifier for matching
2545 strings of non-parentheses is important when applying the pattern to strings
2546 that do not match. For example, when this pattern is applied to
2547 .sp
2548 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
2549 .sp
2550 it yields "no match" quickly. However, if a possessive quantifier is not used,
2551 the match runs for a very long time indeed because there are so many different
2552 ways the + and * repeats can carve up the subject, and all have to be tested
2553 before failure can be reported.
2554 .P
2555 At the end of a match, the values of capturing parentheses are those from
2556 the outermost level. If you want to obtain intermediate values, a callout
2557 function can be used (see below and the
2558 .\" HREF
2559 \fBpcrecallout\fP
2560 .\"
2561 documentation). If the pattern above is matched against
2562 .sp
2563 (ab(cd)ef)
2564 .sp
2565 the value for the inner capturing parentheses (numbered 2) is "ef", which is
2566 the last value taken on at the top level. If a capturing subpattern is not
2567 matched at the top level, its final captured value is unset, even if it was
2568 (temporarily) set at a deeper level during the matching process.
2569 .P
2570 If there are more than 15 capturing parentheses in a pattern, PCRE has to
2571 obtain extra memory to store data during a recursion, which it does by using
2572 \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no memory can
2573 be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
2574 .P
2575 Do not confuse the (?R) item with the condition (R), which tests for recursion.
2576 Consider this pattern, which matches text in angle brackets, allowing for
2577 arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2578 recursing), whereas any characters are permitted at the outer level.
2579 .sp
2580 < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
2581 .sp
2582 In this pattern, (?(R) is the start of a conditional subpattern, with two
2583 different alternatives for the recursive and non-recursive cases. The (?R) item
2584 is the actual recursive call.
2585 .
2586 .
2587 .\" HTML <a name="recursiondifference"></a>
2588 .SS "Differences in recursion processing between PCRE and Perl"
2589 .rs
2590 .sp
2591 Recursion processing in PCRE differs from Perl in two important ways. In PCRE
2592 (like Python, but unlike Perl), a recursive subpattern call is always treated
2593 as an atomic group. That is, once it has matched some of the subject string, it
2594 is never re-entered, even if it contains untried alternatives and there is a
2595 subsequent matching failure. This can be illustrated by the following pattern,
2596 which purports to match a palindromic string that contains an odd number of
2597 characters (for example, "a", "aba", "abcba", "abcdcba"):
2598 .sp
2599 ^(.|(.)(?1)\e2)$
2600 .sp
2601 The idea is that it either matches a single character, or two identical
2602 characters surrounding a sub-palindrome. In Perl, this pattern works; in PCRE
2603 it does not if the pattern is longer than three characters. Consider the
2604 subject string "abcba":
2605 .P
2606 At the top level, the first character is matched, but as it is not at the end
2607 of the string, the first alternative fails; the second alternative is taken
2608 and the recursion kicks in. The recursive call to subpattern 1 successfully
2609 matches the next character ("b"). (Note that the beginning and end of line
2610 tests are not part of the recursion).
2611 .P
2612 Back at the top level, the next character ("c") is compared with what
2613 subpattern 2 matched, which was "a". This fails. Because the recursion is
2614 treated as an atomic group, there are now no backtracking points, and so the
2615 entire match fails. (Perl is able, at this point, to re-enter the recursion and
2616 try the second alternative.) However, if the pattern is written with the
2617 alternatives in the other order, things are different:
2618 .sp
2619 ^((.)(?1)\e2|.)$
2620 .sp
2621 This time, the recursing alternative is tried first, and continues to recurse
2622 until it runs out of characters, at which point the recursion fails. But this
2623 time we do have another alternative to try at the higher level. That is the big
2624 difference: in the previous case the remaining alternative is at a deeper
2625 recursion level, which PCRE cannot use.
2626 .P
2627 To change the pattern so that it matches all palindromic strings, not just
2628 those with an odd number of characters, it is tempting to change the pattern to
2629 this:
2630 .sp
2631 ^((.)(?1)\e2|.?)$
2632 .sp
2633 Again, this works in Perl, but not in PCRE, and for the same reason. When a
2634 deeper recursion has matched a single character, it cannot be entered again in
2635 order to match an empty string. The solution is to separate the two cases, and
2636 write out the odd and even cases as alternatives at the higher level:
2637 .sp
2638 ^(?:((.)(?1)\e2|)|((.)(?3)\e4|.))
2639 .sp
2640 If you want to match typical palindromic phrases, the pattern has to ignore all
2641 non-word characters, which can be done like this:
2642 .sp
2643 ^\eW*+(?:((.)\eW*+(?1)\eW*+\e2|)|((.)\eW*+(?3)\eW*+\e4|\eW*+.\eW*+))\eW*+$
2644 .sp
2645 If run with the PCRE_CASELESS option, this pattern matches phrases such as "A
2646 man, a plan, a canal: Panama!" and it works well in both PCRE and Perl. Note
2647 the use of the possessive quantifier *+ to avoid backtracking into sequences of
2648 non-word characters. Without this, PCRE takes a great deal longer (ten times or
2649 more) to match typical phrases, and Perl takes so long that you think it has
2650 gone into a loop.
2651 .P
2652 \fBWARNING\fP: The palindrome-matching patterns above work only if the subject
2653 string does not start with a palindrome that is shorter than the entire string.
2654 For example, although "abcba" is correctly matched, if the subject is "ababa",
2655 PCRE finds the palindrome "aba" at the start, then fails at top level because
2656 the end of the string does not follow. Once again, it cannot jump back into the
2657 recursion to try other alternatives, so the entire match fails.
2658 .P
2659 The second way in which PCRE and Perl differ in their recursion processing is
2660 in the handling of captured values. In Perl, when a subpattern is called
2661 recursively or as a subpattern (see the next section), it has no access to any
2662 values that were captured outside the recursion, whereas in PCRE these values
2663 can be referenced. Consider this pattern:
2664 .sp
2665 ^(.)(\e1|a(?2))
2666 .sp
2667 In PCRE, this pattern matches "bab". The first capturing parentheses match "b",
2668 then in the second group, when the back reference \e1 fails to match "b", the
2669 second alternative matches "a" and then recurses. In the recursion, \e1 does
2670 now match "b" and so the whole match succeeds. In Perl, the pattern fails to
2671 match because inside the recursive call \e1 cannot access the externally set
2672 value.
2673 .
2674 .
2675 .\" HTML <a name="subpatternsassubroutines"></a>
2677 .rs
2678 .sp
2679 If the syntax for a recursive subpattern call (either by number or by
2680 name) is used outside the parentheses to which it refers, it operates like a
2681 subroutine in a programming language. The called subpattern may be defined
2682 before or after the reference. A numbered reference can be absolute or
2683 relative, as in these examples:
2684 .sp
2685 (...(absolute)...)...(?2)...
2686 (...(relative)...)...(?-1)...
2687 (...(?+1)...(relative)...
2688 .sp
2689 An earlier example pointed out that the pattern
2690 .sp
2691 (sens|respons)e and \e1ibility
2692 .sp
2693 matches "sense and sensibility" and "response and responsibility", but not
2694 "sense and responsibility". If instead the pattern
2695 .sp
2696 (sens|respons)e and (?1)ibility
2697 .sp
2698 is used, it does match "sense and responsibility" as well as the other two
2699 strings. Another example is given in the discussion of DEFINE above.
2700 .P
2701 All subroutine calls, whether recursive or not, are always treated as atomic
2702 groups. That is, once a subroutine has matched some of the subject string, it
2703 is never re-entered, even if it contains untried alternatives and there is a
2704 subsequent matching failure. Any capturing parentheses that are set during the
2705 subroutine call revert to their previous values afterwards.
2706 .P
2707 Processing options such as case-independence are fixed when a subpattern is
2708 defined, so if it is used as a subroutine, such options cannot be changed for
2709 different calls. For example, consider this pattern:
2710 .sp
2711 (abc)(?i:(?-1))
2712 .sp
2713 It matches "abcabc". It does not match "abcABC" because the change of
2714 processing option does not affect the called subpattern.
2715 .
2716 .
2717 .\" HTML <a name="onigurumasubroutines"></a>
2719 .rs
2720 .sp
2721 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
2722 a number enclosed either in angle brackets or single quotes, is an alternative
2723 syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2724 are two of the examples used above, rewritten using this syntax:
2725 .sp
2726 (?<pn> \e( ( (?>[^()]+) | \eg<pn> )* \e) )
2727 (sens|respons)e and \eg'1'ibility
2728 .sp
2729 PCRE supports an extension to Oniguruma: if a number is preceded by a
2730 plus or a minus sign it is taken as a relative reference. For example:
2731 .sp
2732 (abc)(?i:\eg<-1>)
2733 .sp
2734 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
2735 synonymous. The former is a back reference; the latter is a subroutine call.
2736 .
2737 .
2739 .rs
2740 .sp
2741 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2742 code to be obeyed in the middle of matching a regular expression. This makes it
2743 possible, amongst other things, to extract different substrings that match the
2744 same pair of parentheses when there is a repetition.
2745 .P
2746 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2747 code. The feature is called "callout". The caller of PCRE provides an external
2748 function by putting its entry point in the global variable \fIpcre_callout\fP
2749 (8-bit library) or \fIpcre[16|32]_callout\fP (16-bit or 32-bit library).
2750 By default, this variable contains NULL, which disables all calling out.
2751 .P
2752 Within a regular expression, (?C) indicates the points at which the external
2753 function is to be called. If you want to identify different callout points, you
2754 can put a number less than 256 after the letter C. The default value is zero.
2755 For example, this pattern has two callout points:
2756 .sp
2757 (?C1)abc(?C2)def
2758 .sp
2759 If the PCRE_AUTO_CALLOUT flag is passed to a compiling function, callouts are
2760 automatically installed before each item in the pattern. They are all numbered
2761 255. If there is a conditional group in the pattern whose condition is an
2762 assertion, an additional callout is inserted just before the condition. An
2763 explicit callout may also be set at this position, as in this example:
2764 .sp
2765 (?(?C9)(?=a)abc|def)
2766 .sp
2767 Note that this applies only to assertion conditions, not to other types of
2768 condition.
2769 .P
2770 During matching, when PCRE reaches a callout point, the external function is
2771 called. It is provided with the number of the callout, the position in the
2772 pattern, and, optionally, one item of data originally supplied by the caller of
2773 the matching function. The callout function may cause matching to proceed, to
2774 backtrack, or to fail altogether.
2775 .P
2776 By default, PCRE implements a number of optimizations at compile time and
2777 matching time, and one side-effect is that sometimes callouts are skipped. If
2778 you need all possible callouts to happen, you need to set options that disable
2779 the relevant optimizations. More details, and a complete description of the
2780 interface to the callout function, are given in the
2781 .\" HREF
2782 \fBpcrecallout\fP
2783 .\"
2784 documentation.
2785 .
2786 .
2787 .\" HTML <a name="backtrackcontrol"></a>
2789 .rs
2790 .sp
2791 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2792 are still described in the Perl documentation as "experimental and subject to
2793 change or removal in a future version of Perl". It goes on to say: "Their usage
2794 in production code should be noted to avoid problems during upgrades." The same
2795 remarks apply to the PCRE features described in this section.
2796 .P
2797 The new verbs make use of what was previously invalid syntax: an opening
2798 parenthesis followed by an asterisk. They are generally of the form
2799 (*VERB) or (*VERB:NAME). Some may take either form, possibly behaving
2800 differently depending on whether or not a name is present. A name is any
2801 sequence of characters that does not include a closing parenthesis. The maximum
2802 length of name is 255 in the 8-bit library and 65535 in the 16-bit and 32-bit
2803 libraries. If the name is empty, that is, if the closing parenthesis
2804 immediately follows the colon, the effect is as if the colon were not there.
2805 Any number of these verbs may occur in a pattern.
2806 .P
2807 Since these verbs are specifically related to backtracking, most of them can be
2808 used only when the pattern is to be matched using one of the traditional
2809 matching functions, because these use a backtracking algorithm. With the
2810 exception of (*FAIL), which behaves like a failing negative assertion, the
2811 backtracking control verbs cause an error if encountered by a DFA matching
2812 function.
2813 .P
2814 The behaviour of these verbs in
2815 .\" HTML <a href="#btrepeat">
2816 .\" </a>
2817 repeated groups,
2818 .\"
2819 .\" HTML <a href="#btassert">
2820 .\" </a>
2821 assertions,
2822 .\"
2823 and in
2824 .\" HTML <a href="#btsub">
2825 .\" </a>
2826 subpatterns called as subroutines
2827 .\"
2828 (whether or not recursively) is documented below.
2829 .
2830 .
2831 .\" HTML <a name="nooptimize"></a>
2832 .SS "Optimizations that affect backtracking verbs"
2833 .rs
2834 .sp
2835 PCRE contains some optimizations that are used to speed up matching by running
2836 some checks at the start of each match attempt. For example, it may know the
2837 minimum length of matching subject, or that a particular character must be
2838 present. When one of these optimizations bypasses the running of a match, any
2839 included backtracking verbs will not, of course, be processed. You can suppress
2840 the start-of-match optimizations by setting the PCRE_NO_START_OPTIMIZE option
2841 when calling \fBpcre_compile()\fP or \fBpcre_exec()\fP, or by starting the
2842 pattern with (*NO_START_OPT). There is more discussion of this option in the
2843 section entitled
2844 .\" HTML <a href="pcreapi.html#execoptions">
2845 .\" </a>
2846 "Option bits for \fBpcre_exec()\fP"
2847 .\"
2848 in the
2849 .\" HREF
2850 \fBpcreapi\fP
2851 .\"
2852 documentation.
2853 .P
2854 Experiments with Perl suggest that it too has similar optimizations, sometimes
2855 leading to anomalous results.
2856 .
2857 .
2858 .SS "Verbs that act immediately"
2859 .rs
2860 .sp
2861 The following verbs act as soon as they are encountered. They may not be
2862 followed by a name.
2863 .sp
2864 (*ACCEPT)
2865 .sp
2866 This verb causes the match to end successfully, skipping the remainder of the
2867 pattern. However, when it is inside a subpattern that is called as a
2868 subroutine, only that subpattern is ended successfully. Matching then continues
2869 at the outer level. If (*ACCEPT) in triggered in a positive assertion, the
2870 assertion succeeds; in a negative assertion, the assertion fails.
2871 .P
2872 If (*ACCEPT) is inside capturing parentheses, the data so far is captured. For
2873 example:
2874 .sp
2875 A((?:A|B(*ACCEPT)|C)D)
2876 .sp
2877 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is captured by
2878 the outer parentheses.
2879 .sp
2880 (*FAIL) or (*F)
2881 .sp
2882 This verb causes a matching failure, forcing backtracking to occur. It is
2883 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2884 probably useful only when combined with (?{}) or (??{}). Those are, of course,
2885 Perl features that are not present in PCRE. The nearest equivalent is the
2886 callout feature, as for example in this pattern:
2887 .sp
2888 a+(?C)(*FAIL)
2889 .sp
2890 A match with the string "aaaa" always fails, but the callout is taken before
2891 each backtrack happens (in this example, 10 times).
2892 .
2893 .
2894 .SS "Recording which path was taken"
2895 .rs
2896 .sp
2897 There is one verb whose main purpose is to track how a match was arrived at,
2898 though it also has a secondary use in conjunction with advancing the match
2899 starting point (see (*SKIP) below).
2900 .sp
2901 (*MARK:NAME) or (*:NAME)
2902 .sp
2903 A name is always required with this verb. There may be as many instances of
2904 (*MARK) as you like in a pattern, and their names do not have to be unique.
2905 .P
2906 When a match succeeds, the name of the last-encountered (*MARK:NAME),
2907 (*PRUNE:NAME), or (*THEN:NAME) on the matching path is passed back to the
2908 caller as described in the section entitled
2909 .\" HTML <a href="pcreapi.html#extradata">
2910 .\" </a>
2911 "Extra data for \fBpcre_exec()\fP"
2912 .\"
2913 in the
2914 .\" HREF
2915 \fBpcreapi\fP
2916 .\"
2917 documentation. Here is an example of \fBpcretest\fP output, where the /K
2918 modifier requests the retrieval and outputting of (*MARK) data:
2919 .sp
2920 re> /X(*MARK:A)Y|X(*MARK:B)Z/K
2921 data> XY
2922 0: XY
2923 MK: A
2924 XZ
2925 0: XZ
2926 MK: B
2927 .sp
2928 The (*MARK) name is tagged with "MK:" in this output, and in this example it
2929 indicates which of the two alternatives matched. This is a more efficient way
2930 of obtaining this information than putting each alternative in its own
2931 capturing parentheses.
2932 .P
2933 If a verb with a name is encountered in a positive assertion that is true, the
2934 name is recorded and passed back if it is the last-encountered. This does not
2935 happen for negative assertions or failing positive assertions.
2936 .P
2937 After a partial match or a failed match, the last encountered name in the
2938 entire match process is returned. For example:
2939 .sp
2940 re> /X(*MARK:A)Y|X(*MARK:B)Z/K
2941 data> XP
2942 No match, mark = B
2943 .sp
2944 Note that in this unanchored example the mark is retained from the match
2945 attempt that started at the letter "X" in the subject. Subsequent match
2946 attempts starting at "P" and then with an empty string do not get as far as the
2947 (*MARK) item, but nevertheless do not reset it.
2948 .P
2949 If you are interested in (*MARK) values after failed matches, you should
2950 probably set the PCRE_NO_START_OPTIMIZE option
2951 .\" HTML <a href="#nooptimize">
2952 .\" </a>
2953 (see above)
2954 .\"
2955 to ensure that the match is always attempted.
2956 .
2957 .
2958 .SS "Verbs that act after backtracking"
2959 .rs
2960 .sp
2961 The following verbs do nothing when they are encountered. Matching continues
2962 with what follows, but if there is no subsequent match, causing a backtrack to
2963 the verb, a failure is forced. That is, backtracking cannot pass to the left of
2964 the verb. However, when one of these verbs appears inside an atomic group or an
2965 assertion that is true, its effect is confined to that group, because once the
2966 group has been matched, there is never any backtracking into it. In this
2967 situation, backtracking can "jump back" to the left of the entire atomic group
2968 or assertion. (Remember also, as stated above, that this localization also
2969 applies in subroutine calls.)
2970 .P
2971 These verbs differ in exactly what kind of failure occurs when backtracking
2972 reaches them. The behaviour described below is what happens when the verb is
2973 not in a subroutine or an assertion. Subsequent sections cover these special
2974 cases.
2975 .sp
2976 (*COMMIT)
2977 .sp
2978 This verb, which may not be followed by a name, causes the whole match to fail
2979 outright if there is a later matching failure that causes backtracking to reach
2980 it. Even if the pattern is unanchored, no further attempts to find a match by
2981 advancing the starting point take place. If (*COMMIT) is the only backtracking
2982 verb that is encountered, once it has been passed \fBpcre_exec()\fP is
2983 committed to finding a match at the current starting point, or not at all. For
2984 example:
2985 .sp
2986 a+(*COMMIT)b
2987 .sp
2988 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2989 dynamic anchor, or "I've started, so I must finish." The name of the most
2990 recently passed (*MARK) in the path is passed back when (*COMMIT) forces a
2991 match failure.
2992 .P
2993 If there is more than one backtracking verb in a pattern, a different one that
2994 follows (*COMMIT) may be triggered first, so merely passing (*COMMIT) during a
2995 match does not always guarantee that a match must be at this starting point.
2996 .P
2997 Note that (*COMMIT) at the start of a pattern is not the same as an anchor,
2998 unless PCRE's start-of-match optimizations are turned off, as shown in this
2999 \fBpcretest\fP example:
3000 .sp
3001 re> /(*COMMIT)abc/
3002 data> xyzabc
3003 0: abc
3004 xyzabc\eY
3005 No match
3006 .sp
3007 PCRE knows that any match must start with "a", so the optimization skips along
3008 the subject to "a" before running the first match attempt, which succeeds. When
3009 the optimization is disabled by the \eY escape in the second subject, the match
3010 starts at "x" and so the (*COMMIT) causes it to fail without trying any other
3011 starting points.
3012 .sp
3013 (*PRUNE) or (*PRUNE:NAME)
3014 .sp
3015 This verb causes the match to fail at the current starting position in the
3016 subject if there is a later matching failure that causes backtracking to reach
3017 it. If the pattern is unanchored, the normal "bumpalong" advance to the next
3018 starting character then happens. Backtracking can occur as usual to the left of
3019 (*PRUNE), before it is reached, or when matching to the right of (*PRUNE), but
3020 if there is no match to the right, backtracking cannot cross (*PRUNE). In
3021 simple cases, the use of (*PRUNE) is just an alternative to an atomic group or
3022 possessive quantifier, but there are some uses of (*PRUNE) that cannot be
3023 expressed in any other way. In an anchored pattern (*PRUNE) has the same effect
3024 as (*COMMIT).
3025 .P
3026 The behaviour of (*PRUNE:NAME) is the not the same as (*MARK:NAME)(*PRUNE).
3027 It is like (*MARK:NAME) in that the name is remembered for passing back to the
3028 caller. However, (*SKIP:NAME) searches only for names set with (*MARK).
3029 .sp
3030 (*SKIP)
3031 .sp
3032 This verb, when given without a name, is like (*PRUNE), except that if the
3033 pattern is unanchored, the "bumpalong" advance is not to the next character,
3034 but to the position in the subject where (*SKIP) was encountered. (*SKIP)
3035 signifies that whatever text was matched leading up to it cannot be part of a
3036 successful match. Consider:
3037 .sp
3038 a+(*SKIP)b
3039 .sp
3040 If the subject is "aaaac...", after the first match attempt fails (starting at
3041 the first character in the string), the starting point skips on to start the
3042 next attempt at "c". Note that a possessive quantifer does not have the same
3043 effect as this example; although it would suppress backtracking during the
3044 first match attempt, the second attempt would start at the second character
3045 instead of skipping on to "c".
3046 .sp
3047 (*SKIP:NAME)
3048 .sp
3049 When (*SKIP) has an associated name, its behaviour is modified. When it is
3050 triggered, the previous path through the pattern is searched for the most
3051 recent (*MARK) that has the same name. If one is found, the "bumpalong" advance
3052 is to the subject position that corresponds to that (*MARK) instead of to where
3053 (*SKIP) was encountered. If no (*MARK) with a matching name is found, the
3054 (*SKIP) is ignored.
3055 .P
3056 Note that (*SKIP:NAME) searches only for names set by (*MARK:NAME). It ignores
3057 names that are set by (*PRUNE:NAME) or (*THEN:NAME).
3058 .sp
3059 (*THEN) or (*THEN:NAME)
3060 .sp
3061 This verb causes a skip to the next innermost alternative when backtracking
3062 reaches it. That is, it cancels any further backtracking within the current
3063 alternative. Its name comes from the observation that it can be used for a
3064 pattern-based if-then-else block:
3065 .sp
3066 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
3067 .sp
3068 If the COND1 pattern matches, FOO is tried (and possibly further items after
3069 the end of the group if FOO succeeds); on failure, the matcher skips to the
3070 second alternative and tries COND2, without backtracking into COND1. If that
3071 succeeds and BAR fails, COND3 is tried. If subsequently BAZ fails, there are no
3072 more alternatives, so there is a backtrack to whatever came before the entire
3073 group. If (*THEN) is not inside an alternation, it acts like (*PRUNE).
3074 .P
3075 The behaviour of (*THEN:NAME) is the not the same as (*MARK:NAME)(*THEN).
3076 It is like (*MARK:NAME) in that the name is remembered for passing back to the
3077 caller. However, (*SKIP:NAME) searches only for names set with (*MARK).
3078 .P
3079 A subpattern that does not contain a | character is just a part of the
3080 enclosing alternative; it is not a nested alternation with only one
3081 alternative. The effect of (*THEN) extends beyond such a subpattern to the
3082 enclosing alternative. Consider this pattern, where A, B, etc. are complex
3083 pattern fragments that do not contain any | characters at this level:
3084 .sp
3085 A (B(*THEN)C) | D
3086 .sp
3087 If A and B are matched, but there is a failure in C, matching does not
3088 backtrack into A; instead it moves to the next alternative, that is, D.
3089 However, if the subpattern containing (*THEN) is given an alternative, it
3090 behaves differently:
3091 .sp
3092 A (B(*THEN)C | (*FAIL)) | D
3093 .sp
3094 The effect of (*THEN) is now confined to the inner subpattern. After a failure
3095 in C, matching moves to (*FAIL), which causes the whole subpattern to fail
3096 because there are no more alternatives to try. In this case, matching does now
3097 backtrack into A.
3098 .P
3099 Note that a conditional subpattern is not considered as having two
3100 alternatives, because only one is ever used. In other words, the | character in
3101 a conditional subpattern has a different meaning. Ignoring white space,
3102 consider:
3103 .sp
3104 ^.*? (?(?=a) a | b(*THEN)c )
3105 .sp
3106 If the subject is "ba", this pattern does not match. Because .*? is ungreedy,
3107 it initially matches zero characters. The condition (?=a) then fails, the
3108 character "b" is matched, but "c" is not. At this point, matching does not
3109 backtrack to .*? as might perhaps be expected from the presence of the |
3110 character. The conditional subpattern is part of the single alternative that
3111 comprises the whole pattern, and so the match fails. (If there was a backtrack
3112 into .*?, allowing it to match "b", the match would succeed.)
3113 .P
3114 The verbs just described provide four different "strengths" of control when
3115 subsequent matching fails. (*THEN) is the weakest, carrying on the match at the
3116 next alternative. (*PRUNE) comes next, failing the match at the current
3117 starting position, but allowing an advance to the next character (for an
3118 unanchored pattern). (*SKIP) is similar, except that the advance may be more
3119 than one character. (*COMMIT) is the strongest, causing the entire match to
3120 fail.
3121 .
3122 .
3123 .SS "More than one backtracking verb"
3124 .rs
3125 .sp
3126 If more than one backtracking verb is present in a pattern, the one that is
3127 backtracked onto first acts. For example, consider this pattern, where A, B,
3128 etc. are complex pattern fragments:
3129 .sp
3131 .sp
3132 If A matches but B fails, the backtrack to (*COMMIT) causes the entire match to
3133 fail. However, if A and B match, but C fails, the backtrack to (*THEN) causes
3134 the next alternative (ABD) to be tried. This behaviour is consistent, but is
3135 not always the same as Perl's. It means that if two or more backtracking verbs
3136 appear in succession, all the the last of them has no effect. Consider this
3137 example:
3138 .sp
3139 ...(*COMMIT)(*PRUNE)...
3140 .sp
3141 If there is a matching failure to the right, backtracking onto (*PRUNE) causes
3142 it to be triggered, and its action is taken. There can never be a backtrack
3143 onto (*COMMIT).
3144 .
3145 .
3146 .\" HTML <a name="btrepeat"></a>
3147 .SS "Backtracking verbs in repeated groups"
3148 .rs
3149 .sp
3150 PCRE differs from Perl in its handling of backtracking verbs in repeated
3151 groups. For example, consider:
3152 .sp
3153 /(a(*COMMIT)b)+ac/
3154 .sp
3155 If the subject is "abac", Perl matches, but PCRE fails because the (*COMMIT) in
3156 the second repeat of the group acts.
3157 .
3158 .
3159 .\" HTML <a name="btassert"></a>
3160 .SS "Backtracking verbs in assertions"
3161 .rs
3162 .sp
3163 (*FAIL) in an assertion has its normal effect: it forces an immediate backtrack.
3164 .P
3165 (*ACCEPT) in a positive assertion causes the assertion to succeed without any
3166 further processing. In a negative assertion, (*ACCEPT) causes the assertion to
3167 fail without any further processing.
3168 .P
3169 The other backtracking verbs are not treated specially if they appear in a
3170 positive assertion. In particular, (*THEN) skips to the next alternative in the
3171 innermost enclosing group that has alternations, whether or not this is within
3172 the assertion.
3173 .P
3174 Negative assertions are, however, different, in order to ensure that changing a
3175 positive assertion into a negative assertion changes its result. Backtracking
3176 into (*COMMIT), (*SKIP), or (*PRUNE) causes a negative assertion to be true,
3177 without considering any further alternative branches in the assertion.
3178 Backtracking into (*THEN) causes it to skip to the next enclosing alternative
3179 within the assertion (the normal behaviour), but if the assertion does not have
3180 such an alternative, (*THEN) behaves like (*PRUNE).
3181 .
3182 .
3183 .\" HTML <a name="btsub"></a>
3184 .SS "Backtracking verbs in subroutines"
3185 .rs
3186 .sp
3187 These behaviours occur whether or not the subpattern is called recursively.
3188 Perl's treatment of subroutines is different in some cases.
3189 .P
3190 (*FAIL) in a subpattern called as a subroutine has its normal effect: it forces
3191 an immediate backtrack.
3192 .P
3193 (*ACCEPT) in a subpattern called as a subroutine causes the subroutine match to
3194 succeed without any further processing. Matching then continues after the
3195 subroutine call.
3196 .P
3197 (*COMMIT), (*SKIP), and (*PRUNE) in a subpattern called as a subroutine cause
3198 the subroutine match to fail.
3199 .P
3200 (*THEN) skips to the next alternative in the innermost enclosing group within
3201 the subpattern that has alternatives. If there is no such group within the
3202 subpattern, (*THEN) causes the subroutine match to fail.
3203 .
3204 .
3205 .SH "SEE ALSO"
3206 .rs
3207 .sp
3208 \fBpcreapi\fP(3), \fBpcrecallout\fP(3), \fBpcrematching\fP(3),
3209 \fBpcresyntax\fP(3), \fBpcre\fP(3), \fBpcre16(3)\fP, \fBpcre32(3)\fP.
3210 .
3211 .
3213 .rs
3214 .sp
3215 .nf
3216 Philip Hazel
3217 University Computing Service
3218 Cambridge CB2 3QH, England.
3219 .fi
3220 .
3221 .
3223 .rs
3224 .sp
3225 .nf
3226 Last updated: 09 November 2013
3227 Copyright (c) 1997-2013 University of Cambridge.
3228 .fi


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

  ViewVC Help
Powered by ViewVC 1.1.5