ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 556 - (hide annotations) (download)
Tue Oct 26 11:06:44 2010 UTC (4 years, 5 months ago) by ph10
File size: 111333 byte(s)
Fix #-comment bugs in UTF-8 mode with PCRE_NEWLINE_ANY.

1 nigel 79 .TH PCREPATTERN 3
2 nigel 63 .SH NAME
3     PCRE - Perl-compatible regular expressions
5 nigel 63 .rs
6     .sp
7 ph10 208 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 ph10 333 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 ph10 208 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 nigel 75 .P
23     The original operation of PCRE was on strings of one-byte characters. However,
24 ph10 461 there is now also support for UTF-8 character strings. To use this,
25 ph10 456 PCRE must be built to include UTF-8 support, and you must call
26     \fBpcre_compile()\fP or \fBpcre_compile2()\fP with the PCRE_UTF8 option. There
27     is also a special sequence that can be given at the start of a pattern:
28 ph10 412 .sp
29     (*UTF8)
30 ph10 416 .sp
31 ph10 412 Starting a pattern with this sequence is equivalent to setting the PCRE_UTF8
32     option. This feature is not Perl-compatible. How setting UTF-8 mode affects
33     pattern matching is mentioned in several places below. There is also a summary
34     of UTF-8 features in the
35 nigel 63 .\" HTML <a href="pcre.html#utf8support">
36     .\" </a>
37     section on UTF-8 support
38     .\"
39     in the main
40     .\" HREF
41 nigel 75 \fBpcre\fP
42 nigel 63 .\"
43     page.
44 nigel 75 .P
45 ph10 535 Another special sequence that may appear at the start of a pattern or in
46 ph10 518 combination with (*UTF8) is:
47     .sp
48     (*UCP)
49     .sp
50 ph10 535 This has the same effect as setting the PCRE_UCP option: it causes sequences
51     such as \ed and \ew to use Unicode properties to determine character types,
52     instead of recognizing only characters with codes less than 128 via a lookup
53 ph10 518 table.
54     .P
55 nigel 77 The remainder of this document discusses the patterns that are supported by
56     PCRE when its main matching function, \fBpcre_exec()\fP, is used.
57     From release 6.0, PCRE offers a second matching function,
58     \fBpcre_dfa_exec()\fP, which matches using a different algorithm that is not
59 ph10 172 Perl-compatible. Some of the features discussed below are not available when
60 ph10 168 \fBpcre_dfa_exec()\fP is used. The advantages and disadvantages of the
61     alternative function, and how it differs from the normal function, are
62     discussed in the
63 nigel 77 .\" HREF
64     \fBpcrematching\fP
65     .\"
66     page.
67 nigel 93 .
68     .
69 ph10 556 .\" HTML <a name="newlines"></a>
71     .rs
72     .sp
73     PCRE supports five different conventions for indicating line breaks in
74     strings: a single CR (carriage return) character, a single LF (linefeed)
75     character, the two-character sequence CRLF, any of the three preceding, or any
76     Unicode newline sequence. The
77     .\" HREF
78     \fBpcreapi\fP
79     .\"
80     page has
81     .\" HTML <a href="pcreapi.html#newlines">
82     .\" </a>
83     further discussion
84     .\"
85     about newlines, and shows how to set the newline convention in the
86     \fIoptions\fP arguments for the compiling and matching functions.
87     .P
88     It is also possible to specify a newline convention by starting a pattern
89     string with one of the following five sequences:
90     .sp
91     (*CR) carriage return
92     (*LF) linefeed
93     (*CRLF) carriage return, followed by linefeed
94     (*ANYCRLF) any of the three above
95     (*ANY) all Unicode newline sequences
96     .sp
97 ph10 461 These override the default and the options given to \fBpcre_compile()\fP or
98 ph10 456 \fBpcre_compile2()\fP. For example, on a Unix system where LF is the default
99     newline sequence, the pattern
100 ph10 227 .sp
101     (*CR)a.b
102     .sp
103     changes the convention to CR. That pattern matches "a\enb" because LF is no
104     longer a newline. Note that these special settings, which are not
105     Perl-compatible, are recognized only at the very start of a pattern, and that
106 ph10 231 they must be in upper case. If more than one of them is present, the last one
107     is used.
108     .P
109 ph10 514 The newline convention affects the interpretation of the dot metacharacter when
110     PCRE_DOTALL is not set, and also the behaviour of \eN. However, it does not
111     affect what the \eR escape sequence matches. By default, this is any Unicode
112     newline sequence, for Perl compatibility. However, this can be changed; see the
113     description of \eR in the section entitled
114 ph10 231 .\" HTML <a href="#newlineseq">
115     .\" </a>
116     "Newline sequences"
117     .\"
118 ph10 247 below. A change of \eR setting can be combined with a change of newline
119 ph10 246 convention.
120 ph10 227 .
121     .
123     .rs
124     .sp
125 nigel 63 A regular expression is a pattern that is matched against a subject string from
126     left to right. Most characters stand for themselves in a pattern, and match the
127     corresponding characters in the subject. As a trivial example, the pattern
128 nigel 75 .sp
129 nigel 63 The quick brown fox
130 nigel 75 .sp
131 nigel 77 matches a portion of a subject string that is identical to itself. When
132     caseless matching is specified (the PCRE_CASELESS option), letters are matched
133     independently of case. In UTF-8 mode, PCRE always understands the concept of
134     case for characters whose values are less than 128, so caseless matching is
135     always possible. For characters with higher values, the concept of case is
136     supported if PCRE is compiled with Unicode property support, but not otherwise.
137     If you want to use caseless matching for characters 128 and above, you must
138     ensure that PCRE is compiled with Unicode property support as well as with
139     UTF-8 support.
140     .P
141     The power of regular expressions comes from the ability to include alternatives
142     and repetitions in the pattern. These are encoded in the pattern by the use of
143 nigel 75 \fImetacharacters\fP, which do not stand for themselves but instead are
144 nigel 63 interpreted in some special way.
145 nigel 75 .P
146     There are two different sets of metacharacters: those that are recognized
147 nigel 63 anywhere in the pattern except within square brackets, and those that are
148 nigel 93 recognized within square brackets. Outside square brackets, the metacharacters
149     are as follows:
150 nigel 75 .sp
151     \e general escape character with several uses
152 nigel 63 ^ assert start of string (or line, in multiline mode)
153     $ assert end of string (or line, in multiline mode)
154     . match any character except newline (by default)
155     [ start character class definition
156     | start of alternative branch
157     ( start subpattern
158     ) end subpattern
159     ? extends the meaning of (
160     also 0 or 1 quantifier
161     also quantifier minimizer
162     * 0 or more quantifier
163     + 1 or more quantifier
164     also "possessive quantifier"
165     { start min/max quantifier
166 nigel 75 .sp
167 nigel 63 Part of a pattern that is in square brackets is called a "character class". In
168 nigel 75 a character class the only metacharacters are:
169     .sp
170     \e general escape character
171 nigel 63 ^ negate the class, but only if the first character
172     - indicates character range
173 nigel 75 .\" JOIN
174 nigel 63 [ POSIX character class (only if followed by POSIX
175     syntax)
176     ] terminates the character class
177 nigel 75 .sp
178     The following sections describe the use of each of the metacharacters.
179     .
180 nigel 93 .
181 nigel 63 .SH BACKSLASH
182     .rs
183     .sp
184     The backslash character has several uses. Firstly, if it is followed by a
185 nigel 91 non-alphanumeric character, it takes away any special meaning that character
186     may have. This use of backslash as an escape character applies both inside and
187 nigel 63 outside character classes.
188 nigel 75 .P
189     For example, if you want to match a * character, you write \e* in the pattern.
190 nigel 63 This escaping action applies whether or not the following character would
191 nigel 75 otherwise be interpreted as a metacharacter, so it is always safe to precede a
192     non-alphanumeric with backslash to specify that it stands for itself. In
193     particular, if you want to match a backslash, you write \e\e.
194     .P
195 nigel 63 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
196     pattern (other than in a character class) and characters between a # outside
197 nigel 91 a character class and the next newline are ignored. An escaping backslash can
198     be used to include a whitespace or # character as part of the pattern.
199 nigel 75 .P
200 nigel 63 If you want to remove the special meaning from a sequence of characters, you
201 nigel 75 can do so by putting them between \eQ and \eE. This is different from Perl in
202     that $ and @ are handled as literals in \eQ...\eE sequences in PCRE, whereas in
203 nigel 63 Perl, $ and @ cause variable interpolation. Note the following examples:
204 nigel 75 .sp
205 nigel 63 Pattern PCRE matches Perl matches
206 nigel 75 .sp
207     .\" JOIN
208     \eQabc$xyz\eE abc$xyz abc followed by the
209 nigel 63 contents of $xyz
210 nigel 75 \eQabc\e$xyz\eE abc\e$xyz abc\e$xyz
211     \eQabc\eE\e$\eQxyz\eE abc$xyz abc$xyz
212     .sp
213     The \eQ...\eE sequence is recognized both inside and outside character classes.
214 ph10 555 An isolated \eE that is not preceded by \eQ is ignored.
215 nigel 75 .
216     .
217     .\" HTML <a name="digitsafterbackslash"></a>
218     .SS "Non-printing characters"
219     .rs
220     .sp
221 nigel 63 A second use of backslash provides a way of encoding non-printing characters
222     in patterns in a visible manner. There is no restriction on the appearance of
223     non-printing characters, apart from the binary zero that terminates a pattern,
224 ph10 456 but when a pattern is being prepared by text editing, it is often easier to use
225     one of the following escape sequences than the binary character it represents:
226 nigel 75 .sp
227     \ea alarm, that is, the BEL character (hex 07)
228     \ecx "control-x", where x is any character
229     \ee escape (hex 1B)
230     \ef formfeed (hex 0C)
231 ph10 227 \en linefeed (hex 0A)
232 nigel 75 \er carriage return (hex 0D)
233     \et tab (hex 09)
234 ph10 488 \eddd character with octal code ddd, or back reference
235 nigel 75 \exhh character with hex code hh
236 nigel 87 \ex{hhh..} character with hex code hhh..
237 nigel 75 .sp
238     The precise effect of \ecx is as follows: if x is a lower case letter, it
239 nigel 63 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
240 nigel 75 Thus \ecz becomes hex 1A, but \ec{ becomes hex 3B, while \ec; becomes hex
241 nigel 63 7B.
242 nigel 75 .P
243     After \ex, from zero to two hexadecimal digits are read (letters can be in
244 nigel 87 upper or lower case). Any number of hexadecimal digits may appear between \ex{
245     and }, but the value of the character code must be less than 256 in non-UTF-8
246 ph10 211 mode, and less than 2**31 in UTF-8 mode. That is, the maximum value in
247     hexadecimal is 7FFFFFFF. Note that this is bigger than the largest Unicode code
248     point, which is 10FFFF.
249 nigel 75 .P
250 ph10 211 If characters other than hexadecimal digits appear between \ex{ and }, or if
251     there is no terminating }, this form of escape is not recognized. Instead, the
252     initial \ex will be interpreted as a basic hexadecimal escape, with no
253     following digits, giving a character whose value is zero.
254     .P
255 nigel 63 Characters whose value is less than 256 can be defined by either of the two
256 nigel 87 syntaxes for \ex. There is no difference in the way they are handled. For
257     example, \exdc is exactly the same as \ex{dc}.
258 nigel 75 .P
259 nigel 91 After \e0 up to two further octal digits are read. If there are fewer than two
260     digits, just those that are present are used. Thus the sequence \e0\ex\e07
261     specifies two binary zeros followed by a BEL character (code value 7). Make
262     sure you supply two digits after the initial zero if the pattern character that
263     follows is itself an octal digit.
264 nigel 75 .P
265 nigel 63 The handling of a backslash followed by a digit other than 0 is complicated.
266     Outside a character class, PCRE reads it and any following digits as a decimal
267     number. If the number is less than 10, or if there have been at least that many
268     previous capturing left parentheses in the expression, the entire sequence is
269 nigel 75 taken as a \fIback reference\fP. A description of how this works is given
270     .\" HTML <a href="#backreferences">
271     .\" </a>
272     later,
273     .\"
274     following the discussion of
275     .\" HTML <a href="#subpattern">
276     .\" </a>
277     parenthesized subpatterns.
278     .\"
279     .P
280 nigel 63 Inside a character class, or if the decimal number is greater than 9 and there
281     have not been that many capturing subpatterns, PCRE re-reads up to three octal
282 nigel 93 digits following the backslash, and uses them to generate a data character. Any
283 nigel 91 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
284     character specified in octal must be less than \e400. In UTF-8 mode, values up
285     to \e777 are permitted. For example:
286 nigel 75 .sp
287     \e040 is another way of writing a space
288     .\" JOIN
289     \e40 is the same, provided there are fewer than 40
290 nigel 63 previous capturing subpatterns
291 nigel 75 \e7 is always a back reference
292     .\" JOIN
293     \e11 might be a back reference, or another way of
294 nigel 63 writing a tab
295 nigel 75 \e011 is always a tab
296     \e0113 is a tab followed by the character "3"
297     .\" JOIN
298     \e113 might be a back reference, otherwise the
299 nigel 63 character with octal code 113
300 nigel 75 .\" JOIN
301     \e377 might be a back reference, otherwise
302 nigel 63 the byte consisting entirely of 1 bits
303 nigel 75 .\" JOIN
304     \e81 is either a back reference, or a binary zero
305 nigel 63 followed by the two characters "8" and "1"
306 nigel 75 .sp
307 nigel 63 Note that octal values of 100 or greater must not be introduced by a leading
308     zero, because no more than three octal digits are ever read.
309 nigel 75 .P
310 nigel 91 All the sequences that define a single character value can be used both inside
311     and outside character classes. In addition, inside a character class, the
312 ph10 513 sequence \eb is interpreted as the backspace character (hex 08). The sequences
313 ph10 514 \eB, \eN, \eR, and \eX are not special inside a character class. Like any other
314 ph10 513 unrecognized escape sequences, they are treated as the literal characters "B",
315 ph10 514 "N", "R", and "X" by default, but cause an error if the PCRE_EXTRA option is
316     set. Outside a character class, these sequences have different meanings.
317 nigel 75 .
318     .
319 nigel 93 .SS "Absolute and relative back references"
320     .rs
321     .sp
322 ph10 208 The sequence \eg followed by an unsigned or a negative number, optionally
323     enclosed in braces, is an absolute or relative back reference. A named back
324     reference can be coded as \eg{name}. Back references are discussed
325 nigel 93 .\" HTML <a href="#backreferences">
326     .\" </a>
327     later,
328     .\"
329     following the discussion of
330     .\" HTML <a href="#subpattern">
331     .\" </a>
332     parenthesized subpatterns.
333     .\"
334     .
335     .
336 ph10 333 .SS "Absolute and relative subroutine calls"
337     .rs
338     .sp
339 ph10 345 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
340     a number enclosed either in angle brackets or single quotes, is an alternative
341     syntax for referencing a subpattern as a "subroutine". Details are discussed
342 ph10 333 .\" HTML <a href="#onigurumasubroutines">
343     .\" </a>
344     later.
345     .\"
346 ph10 345 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
347 ph10 461 synonymous. The former is a back reference; the latter is a
348 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
349     .\" </a>
350     subroutine
351     .\"
352     call.
353 ph10 333 .
354     .
355 ph10 518 .\" HTML <a name="genericchartypes"></a>
356 nigel 75 .SS "Generic character types"
357     .rs
358     .sp
359 ph10 514 Another use of backslash is for specifying generic character types:
360 nigel 75 .sp
361 ph10 182 \ed any decimal digit
362 nigel 75 \eD any character that is not a decimal digit
363 ph10 178 \eh any horizontal whitespace character
364 ph10 182 \eH any character that is not a horizontal whitespace character
365 nigel 75 \es any whitespace character
366     \eS any character that is not a whitespace character
367 ph10 178 \ev any vertical whitespace character
368 ph10 182 \eV any character that is not a vertical whitespace character
369 nigel 75 \ew any "word" character
370     \eW any "non-word" character
371     .sp
372 ph10 535 There is also the single sequence \eN, which matches a non-newline character.
373     This is the same as
374 ph10 514 .\" HTML <a href="#fullstopdot">
375     .\" </a>
376 ph10 535 the "." metacharacter
377 ph10 514 .\"
378     when PCRE_DOTALL is not set.
379 nigel 75 .P
380 ph10 514 Each pair of lower and upper case escape sequences partitions the complete set
381     of characters into two disjoint sets. Any given character matches one, and only
382 ph10 518 one, of each pair. The sequences can appear both inside and outside character
383 nigel 75 classes. They each match one character of the appropriate type. If the current
384 ph10 518 matching point is at the end of the subject string, all of them fail, because
385 nigel 75 there is no character to match.
386     .P
387     For compatibility with Perl, \es does not match the VT character (code 11).
388     This makes it different from the the POSIX "space" class. The \es characters
389 ph10 178 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
390 nigel 91 included in a Perl script, \es may match the VT character. In PCRE, it never
391 ph10 178 does.
392 nigel 75 .P
393 ph10 518 A "word" character is an underscore or any character that is a letter or digit.
394     By default, the definition of letters and digits is controlled by PCRE's
395     low-valued character tables, and may vary if locale-specific matching is taking
396     place (see
397     .\" HTML <a href="pcreapi.html#localesupport">
398     .\" </a>
399     "Locale support"
400     .\"
401     in the
402     .\" HREF
403     \fBpcreapi\fP
404     .\"
405     page). For example, in a French locale such as "fr_FR" in Unix-like systems,
406     or "french" in Windows, some character codes greater than 128 are used for
407     accented letters, and these are then matched by \ew. The use of locales with
408     Unicode is discouraged.
409 ph10 178 .P
410 ph10 518 By default, in UTF-8 mode, characters with values greater than 128 never match
411     \ed, \es, or \ew, and always match \eD, \eS, and \eW. These sequences retain
412     their original meanings from before UTF-8 support was available, mainly for
413 ph10 535 efficiency reasons. However, if PCRE is compiled with Unicode property support,
414 ph10 518 and the PCRE_UCP option is set, the behaviour is changed so that Unicode
415     properties are used to determine character types, as follows:
416     .sp
417     \ed any character that \ep{Nd} matches (decimal digit)
418     \es any character that \ep{Z} matches, plus HT, LF, FF, CR
419     \ew any character that \ep{L} or \ep{N} matches, plus underscore
420     .sp
421     The upper case escapes match the inverse sets of characters. Note that \ed
422 ph10 535 matches only decimal digits, whereas \ew matches any Unicode digit, as well as
423 ph10 518 any Unicode letter, and underscore. Note also that PCRE_UCP affects \eb, and
424     \eB because they are defined in terms of \ew and \eW. Matching these sequences
425     is noticeably slower when PCRE_UCP is set.
426     .P
427 ph10 182 The sequences \eh, \eH, \ev, and \eV are Perl 5.10 features. In contrast to the
428 ph10 518 other sequences, which match only ASCII characters by default, these always
429 ph10 535 match certain high-valued codepoints in UTF-8 mode, whether or not PCRE_UCP is
430 ph10 518 set. The horizontal space characters are:
431 ph10 178 .sp
432     U+0009 Horizontal tab
433     U+0020 Space
434     U+00A0 Non-break space
435     U+1680 Ogham space mark
436     U+180E Mongolian vowel separator
437     U+2000 En quad
438     U+2001 Em quad
439     U+2002 En space
440     U+2003 Em space
441     U+2004 Three-per-em space
442     U+2005 Four-per-em space
443     U+2006 Six-per-em space
444     U+2007 Figure space
445     U+2008 Punctuation space
446     U+2009 Thin space
447     U+200A Hair space
448     U+202F Narrow no-break space
449     U+205F Medium mathematical space
450     U+3000 Ideographic space
451     .sp
452     The vertical space characters are:
453     .sp
454     U+000A Linefeed
455     U+000B Vertical tab
456     U+000C Formfeed
457     U+000D Carriage return
458     U+0085 Next line
459     U+2028 Line separator
460     U+2029 Paragraph separator
461 nigel 75 .
462     .
463 ph10 231 .\" HTML <a name="newlineseq"></a>
464 nigel 93 .SS "Newline sequences"
465     .rs
466     .sp
467 ph10 231 Outside a character class, by default, the escape sequence \eR matches any
468     Unicode newline sequence. This is a Perl 5.10 feature. In non-UTF-8 mode \eR is
469     equivalent to the following:
470 nigel 93 .sp
471     (?>\er\en|\en|\ex0b|\ef|\er|\ex85)
472     .sp
473     This is an example of an "atomic group", details of which are given
474     .\" HTML <a href="#atomicgroup">
475     .\" </a>
476     below.
477     .\"
478     This particular group matches either the two-character sequence CR followed by
479     LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
480     U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
481     line, U+0085). The two-character sequence is treated as a single unit that
482     cannot be split.
483     .P
484     In UTF-8 mode, two additional characters whose codepoints are greater than 255
485     are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
486     Unicode character property support is not needed for these characters to be
487     recognized.
488     .P
489 ph10 231 It is possible to restrict \eR to match only CR, LF, or CRLF (instead of the
490     complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
491 ph10 247 either at compile time or when the pattern is matched. (BSR is an abbrevation
492 ph10 246 for "backslash R".) This can be made the default when PCRE is built; if this is
493     the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
494     It is also possible to specify these settings by starting a pattern string with
495     one of the following sequences:
496 ph10 231 .sp
497     (*BSR_ANYCRLF) CR, LF, or CRLF only
498     (*BSR_UNICODE) any Unicode newline sequence
499     .sp
500 ph10 461 These override the default and the options given to \fBpcre_compile()\fP or
501 ph10 456 \fBpcre_compile2()\fP, but they can be overridden by options given to
502     \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP. Note that these special settings,
503     which are not Perl-compatible, are recognized only at the very start of a
504     pattern, and that they must be in upper case. If more than one of them is
505     present, the last one is used. They can be combined with a change of newline
506 ph10 518 convention; for example, a pattern can start with:
507 ph10 246 .sp
508     (*ANY)(*BSR_ANYCRLF)
509     .sp
510 ph10 518 They can also be combined with the (*UTF8) or (*UCP) special sequences. Inside
511     a character class, \eR is treated as an unrecognized escape sequence, and so
512     matches the letter "R" by default, but causes an error if PCRE_EXTRA is set.
513 nigel 93 .
514     .
515 nigel 75 .\" HTML <a name="uniextseq"></a>
516     .SS Unicode character properties
517     .rs
518     .sp
519     When PCRE is built with Unicode character property support, three additional
520 ph10 184 escape sequences that match characters with specific properties are available.
521     When not in UTF-8 mode, these sequences are of course limited to testing
522     characters whose codepoints are less than 256, but they do work in this mode.
523     The extra escape sequences are:
524 nigel 75 .sp
525 nigel 87 \ep{\fIxx\fP} a character with the \fIxx\fP property
526     \eP{\fIxx\fP} a character without the \fIxx\fP property
527     \eX an extended Unicode sequence
528 nigel 75 .sp
529 nigel 87 The property names represented by \fIxx\fP above are limited to the Unicode
530 ph10 517 script names, the general category properties, "Any", which matches any
531     character (including newline), and some special PCRE properties (described
532 ph10 535 in the
533 ph10 517 .\" HTML <a href="#extraprops">
534     .\" </a>
535 ph10 535 next section).
536 ph10 517 .\"
537     Other Perl properties such as "InMusicalSymbols" are not currently supported by
538     PCRE. Note that \eP{Any} does not match any characters, so always causes a
539     match failure.
540 nigel 75 .P
541 nigel 87 Sets of Unicode characters are defined as belonging to certain scripts. A
542     character from one of these sets can be matched using a script name. For
543     example:
544 nigel 75 .sp
545 nigel 87 \ep{Greek}
546     \eP{Han}
547     .sp
548     Those that are not part of an identified script are lumped together as
549     "Common". The current list of scripts is:
550     .P
551     Arabic,
552     Armenian,
553 ph10 491 Avestan,
554 nigel 93 Balinese,
555 ph10 491 Bamum,
556 nigel 87 Bengali,
557     Bopomofo,
558     Braille,
559     Buginese,
560     Buhid,
561     Canadian_Aboriginal,
562 ph10 491 Carian,
563     Cham,
564 nigel 87 Cherokee,
565     Common,
566     Coptic,
567 nigel 93 Cuneiform,
568 nigel 87 Cypriot,
569     Cyrillic,
570     Deseret,
571     Devanagari,
572 ph10 491 Egyptian_Hieroglyphs,
573 nigel 87 Ethiopic,
574     Georgian,
575     Glagolitic,
576     Gothic,
577     Greek,
578     Gujarati,
579     Gurmukhi,
580     Han,
581     Hangul,
582     Hanunoo,
583     Hebrew,
584     Hiragana,
585 ph10 491 Imperial_Aramaic,
586 nigel 87 Inherited,
587 ph10 491 Inscriptional_Pahlavi,
588     Inscriptional_Parthian,
589     Javanese,
590     Kaithi,
591 nigel 87 Kannada,
592     Katakana,
593 ph10 491 Kayah_Li,
594 nigel 87 Kharoshthi,
595     Khmer,
596     Lao,
597     Latin,
598 ph10 491 Lepcha,
599 nigel 87 Limbu,
600     Linear_B,
601 ph10 491 Lisu,
602     Lycian,
603     Lydian,
604 nigel 87 Malayalam,
605 ph10 491 Meetei_Mayek,
606 nigel 87 Mongolian,
607     Myanmar,
608     New_Tai_Lue,
609 nigel 93 Nko,
610 nigel 87 Ogham,
611     Old_Italic,
612     Old_Persian,
613 ph10 491 Old_South_Arabian,
614     Old_Turkic,
615     Ol_Chiki,
616 nigel 87 Oriya,
617     Osmanya,
618 nigel 93 Phags_Pa,
619     Phoenician,
620 ph10 491 Rejang,
621 nigel 87 Runic,
622 ph10 491 Samaritan,
623     Saurashtra,
624 nigel 87 Shavian,
625     Sinhala,
626 ph10 491 Sundanese,
627 nigel 87 Syloti_Nagri,
628     Syriac,
629     Tagalog,
630     Tagbanwa,
631     Tai_Le,
632 ph10 491 Tai_Tham,
633     Tai_Viet,
634 nigel 87 Tamil,
635     Telugu,
636     Thaana,
637     Thai,
638     Tibetan,
639     Tifinagh,
640     Ugaritic,
641 ph10 491 Vai,
642 nigel 87 Yi.
643     .P
644 ph10 517 Each character has exactly one Unicode general category property, specified by
645     a two-letter abbreviation. For compatibility with Perl, negation can be
646     specified by including a circumflex between the opening brace and the property
647     name. For example, \ep{^Lu} is the same as \eP{Lu}.
648 nigel 87 .P
649     If only one letter is specified with \ep or \eP, it includes all the general
650     category properties that start with that letter. In this case, in the absence
651     of negation, the curly brackets in the escape sequence are optional; these two
652     examples have the same effect:
653     .sp
654 nigel 75 \ep{L}
655     \epL
656     .sp
657 nigel 87 The following general category property codes are supported:
658 nigel 75 .sp
659     C Other
660     Cc Control
661     Cf Format
662     Cn Unassigned
663     Co Private use
664     Cs Surrogate
665     .sp
666     L Letter
667     Ll Lower case letter
668     Lm Modifier letter
669     Lo Other letter
670     Lt Title case letter
671     Lu Upper case letter
672     .sp
673     M Mark
674     Mc Spacing mark
675     Me Enclosing mark
676     Mn Non-spacing mark
677     .sp
678     N Number
679     Nd Decimal number
680     Nl Letter number
681     No Other number
682     .sp
683     P Punctuation
684     Pc Connector punctuation
685     Pd Dash punctuation
686     Pe Close punctuation
687     Pf Final punctuation
688     Pi Initial punctuation
689     Po Other punctuation
690     Ps Open punctuation
691     .sp
692     S Symbol
693     Sc Currency symbol
694     Sk Modifier symbol
695     Sm Mathematical symbol
696     So Other symbol
697     .sp
698     Z Separator
699     Zl Line separator
700     Zp Paragraph separator
701     Zs Space separator
702     .sp
703 nigel 87 The special property L& is also supported: it matches a character that has
704     the Lu, Ll, or Lt property, in other words, a letter that is not classified as
705     a modifier or "other".
706 nigel 75 .P
707 ph10 211 The Cs (Surrogate) property applies only to characters in the range U+D800 to
708     U+DFFF. Such characters are not valid in UTF-8 strings (see RFC 3629) and so
709     cannot be tested by PCRE, unless UTF-8 validity checking has been turned off
710     (see the discussion of PCRE_NO_UTF8_CHECK in the
711     .\" HREF
712     \fBpcreapi\fP
713     .\"
714 ph10 451 page). Perl does not support the Cs property.
715 ph10 211 .P
716 ph10 451 The long synonyms for property names that Perl supports (such as \ep{Letter})
717 nigel 91 are not supported by PCRE, nor is it permitted to prefix any of these
718 nigel 87 properties with "Is".
719     .P
720     No character that is in the Unicode table has the Cn (unassigned) property.
721     Instead, this property is assumed for any code point that is not in the
722     Unicode table.
723     .P
724 nigel 75 Specifying caseless matching does not affect these escape sequences. For
725     example, \ep{Lu} always matches only upper case letters.
726     .P
727     The \eX escape matches any number of Unicode characters that form an extended
728     Unicode sequence. \eX is equivalent to
729     .sp
730     (?>\ePM\epM*)
731     .sp
732     That is, it matches a character without the "mark" property, followed by zero
733     or more characters with the "mark" property, and treats the sequence as an
734     atomic group
735     .\" HTML <a href="#atomicgroup">
736     .\" </a>
737     (see below).
738     .\"
739     Characters with the "mark" property are typically accents that affect the
740 ph10 185 preceding character. None of them have codepoints less than 256, so in
741 ph10 184 non-UTF-8 mode \eX matches any one character.
742 nigel 75 .P
743     Matching characters by Unicode property is not fast, because PCRE has to search
744     a structure that contains data for over fifteen thousand characters. That is
745     why the traditional escape sequences such as \ed and \ew do not use Unicode
746 ph10 535 properties in PCRE by default, though you can make them do so by setting the
747 ph10 518 PCRE_UCP option for \fBpcre_compile()\fP or by starting the pattern with
748     (*UCP).
749 nigel 75 .
750     .
751 ph10 517 .\" HTML <a name="extraprops"></a>
752     .SS PCRE's additional properties
753     .rs
754     .sp
755 ph10 535 As well as the standard Unicode properties described in the previous
756     section, PCRE supports four more that make it possible to convert traditional
757 ph10 517 escape sequences such as \ew and \es and POSIX character classes to use Unicode
758 ph10 518 properties. PCRE uses these non-standard, non-Perl properties internally when
759     PCRE_UCP is set. They are:
760 ph10 517 .sp
761     Xan Any alphanumeric character
762     Xps Any POSIX space character
763     Xsp Any Perl space character
764     Xwd Any Perl "word" character
765     .sp
766 ph10 535 Xan matches characters that have either the L (letter) or the N (number)
767     property. Xps matches the characters tab, linefeed, vertical tab, formfeed, or
768 ph10 517 carriage return, and any other character that has the Z (separator) property.
769 ph10 535 Xsp is the same as Xps, except that vertical tab is excluded. Xwd matches the
770 ph10 517 same characters as Xan, plus underscore.
771     .
772     .
773 ph10 168 .\" HTML <a name="resetmatchstart"></a>
774     .SS "Resetting the match start"
775     .rs
776     .sp
777 ph10 172 The escape sequence \eK, which is a Perl 5.10 feature, causes any previously
778     matched characters not to be included in the final matched sequence. For
779 ph10 168 example, the pattern:
780     .sp
781     foo\eKbar
782     .sp
783 ph10 172 matches "foobar", but reports that it has matched "bar". This feature is
784 ph10 168 similar to a lookbehind assertion
785     .\" HTML <a href="#lookbehind">
786     .\" </a>
787     (described below).
788     .\"
789 ph10 172 However, in this case, the part of the subject before the real match does not
790     have to be of fixed length, as lookbehind assertions do. The use of \eK does
791 ph10 168 not interfere with the setting of
792     .\" HTML <a href="#subpattern">
793     .\" </a>
794     captured substrings.
795 ph10 172 .\"
796 ph10 168 For example, when the pattern
797     .sp
798     (foo)\eKbar
799     .sp
800 ph10 172 matches "foobar", the first substring is still set to "foo".
801 ph10 500 .P
802 ph10 507 Perl documents that the use of \eK within assertions is "not well defined". In
803     PCRE, \eK is acted upon when it occurs inside positive assertions, but is
804 ph10 500 ignored in negative assertions.
805 ph10 168 .
806     .
807 nigel 75 .\" HTML <a name="smallassertions"></a>
808     .SS "Simple assertions"
809     .rs
810     .sp
811 nigel 93 The final use of backslash is for certain simple assertions. An assertion
812 nigel 63 specifies a condition that has to be met at a particular point in a match,
813     without consuming any characters from the subject string. The use of
814 nigel 75 subpatterns for more complicated assertions is described
815     .\" HTML <a href="#bigassertions">
816     .\" </a>
817     below.
818     .\"
819 nigel 91 The backslashed assertions are:
820 nigel 75 .sp
821     \eb matches at a word boundary
822     \eB matches when not at a word boundary
823 nigel 93 \eA matches at the start of the subject
824     \eZ matches at the end of the subject
825     also matches before a newline at the end of the subject
826     \ez matches only at the end of the subject
827     \eG matches at the first matching position in the subject
828 nigel 75 .sp
829 ph10 513 Inside a character class, \eb has a different meaning; it matches the backspace
830 ph10 535 character. If any other of these assertions appears in a character class, by
831 ph10 513 default it matches the corresponding literal character (for example, \eB
832     matches the letter B). However, if the PCRE_EXTRA option is set, an "invalid
833     escape sequence" error is generated instead.
834 nigel 75 .P
835 nigel 63 A word boundary is a position in the subject string where the current character
836 nigel 75 and the previous character do not both match \ew or \eW (i.e. one matches
837     \ew and the other matches \eW), or the start or end of the string if the
838 ph10 518 first or last character matches \ew, respectively. In UTF-8 mode, the meanings
839     of \ew and \eW can be changed by setting the PCRE_UCP option. When this is
840     done, it also affects \eb and \eB. Neither PCRE nor Perl has a separate "start
841     of word" or "end of word" metasequence. However, whatever follows \eb normally
842     determines which it is. For example, the fragment \eba matches "a" at the start
843     of a word.
844 nigel 75 .P
845     The \eA, \eZ, and \ez assertions differ from the traditional circumflex and
846     dollar (described in the next section) in that they only ever match at the very
847     start and end of the subject string, whatever options are set. Thus, they are
848     independent of multiline mode. These three assertions are not affected by the
849     PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
850     circumflex and dollar metacharacters. However, if the \fIstartoffset\fP
851     argument of \fBpcre_exec()\fP is non-zero, indicating that matching is to start
852     at a point other than the beginning of the subject, \eA can never match. The
853 nigel 91 difference between \eZ and \ez is that \eZ matches before a newline at the end
854     of the string as well as at the very end, whereas \ez matches only at the end.
855 nigel 75 .P
856     The \eG assertion is true only when the current matching position is at the
857     start point of the match, as specified by the \fIstartoffset\fP argument of
858     \fBpcre_exec()\fP. It differs from \eA when the value of \fIstartoffset\fP is
859     non-zero. By calling \fBpcre_exec()\fP multiple times with appropriate
860 nigel 63 arguments, you can mimic Perl's /g option, and it is in this kind of
861 nigel 75 implementation where \eG can be useful.
862     .P
863     Note, however, that PCRE's interpretation of \eG, as the start of the current
864 nigel 63 match, is subtly different from Perl's, which defines it as the end of the
865     previous match. In Perl, these can be different when the previously matched
866     string was empty. Because PCRE does just one match at a time, it cannot
867     reproduce this behaviour.
868 nigel 75 .P
869     If all the alternatives of a pattern begin with \eG, the expression is anchored
870 nigel 63 to the starting match position, and the "anchored" flag is set in the compiled
871     regular expression.
872 nigel 75 .
873     .
875 nigel 63 .rs
876     .sp
877     Outside a character class, in the default matching mode, the circumflex
878 nigel 75 character is an assertion that is true only if the current matching point is
879     at the start of the subject string. If the \fIstartoffset\fP argument of
880     \fBpcre_exec()\fP is non-zero, circumflex can never match if the PCRE_MULTILINE
881 nigel 63 option is unset. Inside a character class, circumflex has an entirely different
882 nigel 75 meaning
883     .\" HTML <a href="#characterclass">
884     .\" </a>
885     (see below).
886     .\"
887     .P
888 nigel 63 Circumflex need not be the first character of the pattern if a number of
889     alternatives are involved, but it should be the first thing in each alternative
890     in which it appears if the pattern is ever to match that branch. If all
891     possible alternatives start with a circumflex, that is, if the pattern is
892     constrained to match only at the start of the subject, it is said to be an
893     "anchored" pattern. (There are also other constructs that can cause a pattern
894     to be anchored.)
895 nigel 75 .P
896     A dollar character is an assertion that is true only if the current matching
897 nigel 63 point is at the end of the subject string, or immediately before a newline
898 nigel 91 at the end of the string (by default). Dollar need not be the last character of
899     the pattern if a number of alternatives are involved, but it should be the last
900     item in any branch in which it appears. Dollar has no special meaning in a
901     character class.
902 nigel 75 .P
903 nigel 63 The meaning of dollar can be changed so that it matches only at the very end of
904     the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
905 nigel 75 does not affect the \eZ assertion.
906     .P
907 nigel 63 The meanings of the circumflex and dollar characters are changed if the
908 nigel 91 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
909     immediately after internal newlines as well as at the start of the subject
910     string. It does not match after a newline that ends the string. A dollar
911     matches before any newlines in the string, as well as at the very end, when
912     PCRE_MULTILINE is set. When newline is specified as the two-character
913     sequence CRLF, isolated CR and LF characters do not indicate newlines.
914 nigel 75 .P
915 nigel 91 For example, the pattern /^abc$/ matches the subject string "def\enabc" (where
916     \en represents a newline) in multiline mode, but not otherwise. Consequently,
917     patterns that are anchored in single line mode because all branches start with
918     ^ are not anchored in multiline mode, and a match for circumflex is possible
919     when the \fIstartoffset\fP argument of \fBpcre_exec()\fP is non-zero. The
920     PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
921     .P
922 nigel 75 Note that the sequences \eA, \eZ, and \ez can be used to match the start and
923 nigel 63 end of the subject in both modes, and if all branches of a pattern start with
924 nigel 91 \eA it is always anchored, whether or not PCRE_MULTILINE is set.
925 nigel 75 .
926     .
927 ph10 514 .\" HTML <a name="fullstopdot"></a>
929 nigel 63 .rs
930     .sp
931     Outside a character class, a dot in the pattern matches any one character in
932 nigel 91 the subject string except (by default) a character that signifies the end of a
933 nigel 93 line. In UTF-8 mode, the matched character may be more than one byte long.
934 nigel 91 .P
935 nigel 93 When a line ending is defined as a single character, dot never matches that
936     character; when the two-character sequence CRLF is used, dot does not match CR
937     if it is immediately followed by LF, but otherwise it matches all characters
938     (including isolated CRs and LFs). When any Unicode line endings are being
939     recognized, dot does not match CR or LF or any of the other line ending
940     characters.
941     .P
942 nigel 91 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
943 nigel 93 option is set, a dot matches any one character, without exception. If the
944     two-character sequence CRLF is present in the subject string, it takes two dots
945     to match it.
946 nigel 91 .P
947     The handling of dot is entirely independent of the handling of circumflex and
948     dollar, the only relationship being that they both involve newlines. Dot has no
949     special meaning in a character class.
950 ph10 514 .P
951 ph10 535 The escape sequence \eN always behaves as a dot does when PCRE_DOTALL is not
952     set. In other words, it matches any one character except one that signifies the
953 ph10 514 end of a line.
954 nigel 75 .
955     .
957 nigel 63 .rs
958     .sp
959 nigel 75 Outside a character class, the escape sequence \eC matches any one byte, both
960 nigel 93 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
961     characters. The feature is provided in Perl in order to match individual bytes
962     in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes,
963     what remains in the string may be a malformed UTF-8 string. For this reason,
964     the \eC escape sequence is best avoided.
965 nigel 75 .P
966     PCRE does not allow \eC to appear in lookbehind assertions
967     .\" HTML <a href="#lookbehind">
968     .\" </a>
969     (described below),
970     .\"
971     because in UTF-8 mode this would make it impossible to calculate the length of
972     the lookbehind.
973     .
974     .
975     .\" HTML <a name="characterclass"></a>
977 nigel 63 .rs
978     .sp
979     An opening square bracket introduces a character class, terminated by a closing
980 ph10 461 square bracket. A closing square bracket on its own is not special by default.
981     However, if the PCRE_JAVASCRIPT_COMPAT option is set, a lone closing square
982 ph10 456 bracket causes a compile-time error. If a closing square bracket is required as
983     a member of the class, it should be the first data character in the class
984     (after an initial circumflex, if present) or escaped with a backslash.
985 nigel 75 .P
986 nigel 63 A character class matches a single character in the subject. In UTF-8 mode, the
987 ph10 456 character may be more than one byte long. A matched character must be in the
988     set of characters defined by the class, unless the first character in the class
989 nigel 63 definition is a circumflex, in which case the subject character must not be in
990     the set defined by the class. If a circumflex is actually required as a member
991     of the class, ensure it is not the first character, or escape it with a
992     backslash.
993 nigel 75 .P
994 nigel 63 For example, the character class [aeiou] matches any lower case vowel, while
995     [^aeiou] matches any character that is not a lower case vowel. Note that a
996 nigel 75 circumflex is just a convenient notation for specifying the characters that
997     are in the class by enumerating those that are not. A class that starts with a
998 ph10 456 circumflex is not an assertion; it still consumes a character from the subject
999 nigel 75 string, and therefore it fails if the current pointer is at the end of the
1000     string.
1001     .P
1002 nigel 63 In UTF-8 mode, characters with values greater than 255 can be included in a
1003 nigel 75 class as a literal string of bytes, or by using the \ex{ escaping mechanism.
1004     .P
1005 nigel 63 When caseless matching is set, any letters in a class represent both their
1006     upper case and lower case versions, so for example, a caseless [aeiou] matches
1007     "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
1008 nigel 77 caseful version would. In UTF-8 mode, PCRE always understands the concept of
1009     case for characters whose values are less than 128, so caseless matching is
1010     always possible. For characters with higher values, the concept of case is
1011     supported if PCRE is compiled with Unicode property support, but not otherwise.
1012 ph10 456 If you want to use caseless matching in UTF8-mode for characters 128 and above,
1013     you must ensure that PCRE is compiled with Unicode property support as well as
1014     with UTF-8 support.
1015 nigel 75 .P
1016 nigel 93 Characters that might indicate line breaks are never treated in any special way
1017     when matching character classes, whatever line-ending sequence is in use, and
1018     whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
1019     such as [^a] always matches one of these characters.
1020 nigel 75 .P
1021 nigel 63 The minus (hyphen) character can be used to specify a range of characters in a
1022     character class. For example, [d-m] matches any letter between d and m,
1023     inclusive. If a minus character is required in a class, it must be escaped with
1024     a backslash or appear in a position where it cannot be interpreted as
1025     indicating a range, typically as the first or last character in the class.
1026 nigel 75 .P
1027 nigel 63 It is not possible to have the literal character "]" as the end character of a
1028     range. A pattern such as [W-]46] is interpreted as a class of two characters
1029     ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
1030     "-46]". However, if the "]" is escaped with a backslash it is interpreted as
1031 nigel 75 the end of range, so [W-\e]46] is interpreted as a class containing a range
1032     followed by two other characters. The octal or hexadecimal representation of
1033     "]" can also be used to end a range.
1034     .P
1035 nigel 63 Ranges operate in the collating sequence of character values. They can also be
1036 nigel 75 used for characters specified numerically, for example [\e000-\e037]. In UTF-8
1037 nigel 63 mode, ranges can include characters whose values are greater than 255, for
1038 nigel 75 example [\ex{100}-\ex{2ff}].
1039     .P
1040 nigel 63 If a range that includes letters is used when caseless matching is set, it
1041     matches the letters in either case. For example, [W-c] is equivalent to
1042 nigel 75 [][\e\e^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
1043 ph10 139 tables for a French locale are in use, [\exc8-\excb] matches accented E
1044 nigel 75 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
1045     characters with values greater than 128 only when it is compiled with Unicode
1046     property support.
1047     .P
1048 ph10 518 The character types \ed, \eD, \eh, \eH, \ep, \eP, \es, \eS, \ev, \eV, \ew, and
1049     \eW may also appear in a character class, and add the characters that they
1050     match to the class. For example, [\edABCDEF] matches any hexadecimal digit. A
1051     circumflex can conveniently be used with the upper case character types to
1052     specify a more restricted set of characters than the matching lower case type.
1053     For example, the class [^\eW_] matches any letter or digit, but not underscore.
1054 nigel 75 .P
1055     The only metacharacters that are recognized in character classes are backslash,
1056     hyphen (only where it can be interpreted as specifying a range), circumflex
1057     (only at the start), opening square bracket (only when it can be interpreted as
1058     introducing a POSIX class name - see the next section), and the terminating
1059     closing square bracket. However, escaping other non-alphanumeric characters
1060     does no harm.
1061     .
1062     .
1064 nigel 63 .rs
1065     .sp
1066 nigel 75 Perl supports the POSIX notation for character classes. This uses names
1067 nigel 63 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
1068     this notation. For example,
1069 nigel 75 .sp
1070 nigel 63 [01[:alpha:]%]
1071 nigel 75 .sp
1072 nigel 63 matches "0", "1", any alphabetic character, or "%". The supported class names
1073 ph10 518 are:
1074 nigel 75 .sp
1075 nigel 63 alnum letters and digits
1076     alpha letters
1077     ascii character codes 0 - 127
1078     blank space or tab only
1079     cntrl control characters
1080 nigel 75 digit decimal digits (same as \ed)
1081 nigel 63 graph printing characters, excluding space
1082     lower lower case letters
1083     print printing characters, including space
1084 ph10 518 punct printing characters, excluding letters and digits and space
1085 nigel 75 space white space (not quite the same as \es)
1086 nigel 63 upper upper case letters
1087 nigel 75 word "word" characters (same as \ew)
1088 nigel 63 xdigit hexadecimal digits
1089 nigel 75 .sp
1090 nigel 63 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
1091     space (32). Notice that this list includes the VT character (code 11). This
1092 nigel 75 makes "space" different to \es, which does not include VT (for Perl
1093 nigel 63 compatibility).
1094 nigel 75 .P
1095 nigel 63 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
1096     5.8. Another Perl extension is negation, which is indicated by a ^ character
1097     after the colon. For example,
1098 nigel 75 .sp
1099 nigel 63 [12[:^digit:]]
1100 nigel 75 .sp
1101 nigel 63 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
1102     syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
1103     supported, and an error is given if they are encountered.
1104 nigel 75 .P
1105 ph10 518 By default, in UTF-8 mode, characters with values greater than 128 do not match
1106     any of the POSIX character classes. However, if the PCRE_UCP option is passed
1107 ph10 535 to \fBpcre_compile()\fP, some of the classes are changed so that Unicode
1108     character properties are used. This is achieved by replacing the POSIX classes
1109 ph10 518 by other sequences, as follows:
1110     .sp
1111     [:alnum:] becomes \ep{Xan}
1112     [:alpha:] becomes \ep{L}
1113 ph10 535 [:blank:] becomes \eh
1114 ph10 518 [:digit:] becomes \ep{Nd}
1115     [:lower:] becomes \ep{Ll}
1116 ph10 535 [:space:] becomes \ep{Xps}
1117 ph10 518 [:upper:] becomes \ep{Lu}
1118     [:word:] becomes \ep{Xwd}
1119     .sp
1120     Negated versions, such as [:^alpha:] use \eP instead of \ep. The other POSIX
1121     classes are unchanged, and match only characters with code points less than
1122     128.
1123 nigel 75 .
1124     .
1125     .SH "VERTICAL BAR"
1126 nigel 63 .rs
1127     .sp
1128     Vertical bar characters are used to separate alternative patterns. For example,
1129     the pattern
1130 nigel 75 .sp
1131 nigel 63 gilbert|sullivan
1132 nigel 75 .sp
1133 nigel 63 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
1134 nigel 91 and an empty alternative is permitted (matching the empty string). The matching
1135     process tries each alternative in turn, from left to right, and the first one
1136     that succeeds is used. If the alternatives are within a subpattern
1137 nigel 75 .\" HTML <a href="#subpattern">
1138     .\" </a>
1139     (defined below),
1140     .\"
1141     "succeeds" means matching the rest of the main pattern as well as the
1142     alternative in the subpattern.
1143     .
1144     .
1146 nigel 63 .rs
1147     .sp
1148     The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1149 ph10 231 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1150     the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1151     The option letters are
1152 nigel 75 .sp
1153 nigel 63 i for PCRE_CASELESS
1154     m for PCRE_MULTILINE
1155     s for PCRE_DOTALL
1156     x for PCRE_EXTENDED
1157 nigel 75 .sp
1158 nigel 63 For example, (?im) sets caseless, multiline matching. It is also possible to
1159     unset these options by preceding the letter with a hyphen, and a combined
1160     setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1161     PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1162     permitted. If a letter appears both before and after the hyphen, the option is
1163     unset.
1164 nigel 75 .P
1165 ph10 231 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1166     changed in the same way as the Perl-compatible options by using the characters
1167     J, U and X respectively.
1168     .P
1169 ph10 412 When one of these option changes occurs at top level (that is, not inside
1170     subpattern parentheses), the change applies to the remainder of the pattern
1171     that follows. If the change is placed right at the start of a pattern, PCRE
1172     extracts it into the global options (and it will therefore show up in data
1173     extracted by the \fBpcre_fullinfo()\fP function).
1174 nigel 75 .P
1175 nigel 93 An option change within a subpattern (see below for a description of
1176     subpatterns) affects only that part of the current pattern that follows it, so
1177 nigel 75 .sp
1178 nigel 63 (a(?i)b)c
1179 nigel 75 .sp
1180 nigel 63 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1181     By this means, options can be made to have different settings in different
1182     parts of the pattern. Any changes made in one alternative do carry on
1183     into subsequent branches within the same subpattern. For example,
1184 nigel 75 .sp
1185 nigel 63 (a(?i)b|c)
1186 nigel 75 .sp
1187 nigel 63 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1188     branch is abandoned before the option setting. This is because the effects of
1189     option settings happen at compile time. There would be some very weird
1190     behaviour otherwise.
1191 ph10 251 .P
1192     \fBNote:\fP There are other PCRE-specific options that can be set by the
1193     application when the compile or match functions are called. In some cases the
1194 ph10 412 pattern can contain special leading sequences such as (*CRLF) to override what
1195     the application has set or what has been defaulted. Details are given in the
1196     section entitled
1197 ph10 251 .\" HTML <a href="#newlineseq">
1198     .\" </a>
1199     "Newline sequences"
1200     .\"
1201 ph10 518 above. There are also the (*UTF8) and (*UCP) leading sequences that can be used
1202     to set UTF-8 and Unicode property modes; they are equivalent to setting the
1203     PCRE_UTF8 and the PCRE_UCP options, respectively.
1204 nigel 75 .
1205     .
1206     .\" HTML <a name="subpattern"></a>
1207 nigel 63 .SH SUBPATTERNS
1208     .rs
1209     .sp
1210     Subpatterns are delimited by parentheses (round brackets), which can be nested.
1211 nigel 75 Turning part of a pattern into a subpattern does two things:
1212     .sp
1213 nigel 63 1. It localizes a set of alternatives. For example, the pattern
1214 nigel 75 .sp
1215 nigel 63 cat(aract|erpillar|)
1216 nigel 75 .sp
1217 nigel 63 matches one of the words "cat", "cataract", or "caterpillar". Without the
1218 nigel 93 parentheses, it would match "cataract", "erpillar" or an empty string.
1219 nigel 75 .sp
1220     2. It sets up the subpattern as a capturing subpattern. This means that, when
1221     the whole pattern matches, that portion of the subject string that matched the
1222     subpattern is passed back to the caller via the \fIovector\fP argument of
1223     \fBpcre_exec()\fP. Opening parentheses are counted from left to right (starting
1224     from 1) to obtain numbers for the capturing subpatterns.
1225     .P
1226 nigel 63 For example, if the string "the red king" is matched against the pattern
1227 nigel 75 .sp
1228 nigel 63 the ((red|white) (king|queen))
1229 nigel 75 .sp
1230 nigel 63 the captured substrings are "red king", "red", and "king", and are numbered 1,
1231     2, and 3, respectively.
1232 nigel 75 .P
1233 nigel 63 The fact that plain parentheses fulfil two functions is not always helpful.
1234     There are often times when a grouping subpattern is required without a
1235     capturing requirement. If an opening parenthesis is followed by a question mark
1236     and a colon, the subpattern does not do any capturing, and is not counted when
1237     computing the number of any subsequent capturing subpatterns. For example, if
1238     the string "the white queen" is matched against the pattern
1239 nigel 75 .sp
1240 nigel 63 the ((?:red|white) (king|queen))
1241 nigel 75 .sp
1242 nigel 63 the captured substrings are "white queen" and "queen", and are numbered 1 and
1243 nigel 93 2. The maximum number of capturing subpatterns is 65535.
1244 nigel 75 .P
1245 nigel 63 As a convenient shorthand, if any option settings are required at the start of
1246     a non-capturing subpattern, the option letters may appear between the "?" and
1247     the ":". Thus the two patterns
1248 nigel 75 .sp
1249 nigel 63 (?i:saturday|sunday)
1250     (?:(?i)saturday|sunday)
1251 nigel 75 .sp
1252 nigel 63 match exactly the same set of strings. Because alternative branches are tried
1253     from left to right, and options are not reset until the end of the subpattern
1254     is reached, an option setting in one branch does affect subsequent branches, so
1255     the above patterns match "SUNDAY" as well as "Saturday".
1256 nigel 75 .
1257     .
1258 ph10 456 .\" HTML <a name="dupsubpatternnumber"></a>
1260     .rs
1261     .sp
1262 ph10 182 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1263     the same numbers for its capturing parentheses. Such a subpattern starts with
1264     (?| and is itself a non-capturing subpattern. For example, consider this
1265 ph10 175 pattern:
1266     .sp
1267     (?|(Sat)ur|(Sun))day
1268 ph10 182 .sp
1269     Because the two alternatives are inside a (?| group, both sets of capturing
1270     parentheses are numbered one. Thus, when the pattern matches, you can look
1271     at captured substring number one, whichever alternative matched. This construct
1272     is useful when you want to capture part, but not all, of one of a number of
1273     alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1274 ph10 175 number is reset at the start of each branch. The numbers of any capturing
1275 ph10 182 buffers that follow the subpattern start after the highest number used in any
1276     branch. The following example is taken from the Perl documentation.
1277 ph10 175 The numbers underneath show in which buffer the captured content will be
1278     stored.
1279     .sp
1280     # before ---------------branch-reset----------- after
1281     / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1282     # 1 2 2 3 2 3 4
1283 ph10 182 .sp
1284 ph10 488 A back reference to a numbered subpattern uses the most recent value that is
1285     set for that number by any subpattern. The following pattern matches "abcabc"
1286     or "defdef":
1287 ph10 456 .sp
1288 ph10 461 /(?|(abc)|(def))\e1/
1289 ph10 456 .sp
1290     In contrast, a recursive or "subroutine" call to a numbered subpattern always
1291 ph10 461 refers to the first one in the pattern with the given number. The following
1292 ph10 456 pattern matches "abcabc" or "defabc":
1293     .sp
1294     /(?|(abc)|(def))(?1)/
1295     .sp
1296 ph10 459 If a
1297     .\" HTML <a href="#conditions">
1298     .\" </a>
1299     condition test
1300     .\"
1301     for a subpattern's having matched refers to a non-unique number, the test is
1302     true if any of the subpatterns of that number have matched.
1303     .P
1304     An alternative approach to using this "branch reset" feature is to use
1305 ph10 175 duplicate named subpatterns, as described in the next section.
1306     .
1307     .
1308 nigel 75 .SH "NAMED SUBPATTERNS"
1309 nigel 63 .rs
1310     .sp
1311     Identifying capturing parentheses by number is simple, but it can be very hard
1312     to keep track of the numbers in complicated regular expressions. Furthermore,
1313 nigel 75 if an expression is modified, the numbers may change. To help with this
1314 nigel 93 difficulty, PCRE supports the naming of subpatterns. This feature was not
1315     added to Perl until release 5.10. Python had the feature earlier, and PCRE
1316     introduced it at release 4.0, using the Python syntax. PCRE now supports both
1317 ph10 459 the Perl and the Python syntax. Perl allows identically numbered subpatterns to
1318     have different names, but PCRE does not.
1319 nigel 93 .P
1320     In PCRE, a subpattern can be named in one of three ways: (?<name>...) or
1321     (?'name'...) as in Perl, or (?P<name>...) as in Python. References to capturing
1322 nigel 91 parentheses from other parts of the pattern, such as
1323     .\" HTML <a href="#backreferences">
1324     .\" </a>
1325 ph10 488 back references,
1326 nigel 91 .\"
1327     .\" HTML <a href="#recursion">
1328     .\" </a>
1329     recursion,
1330     .\"
1331     and
1332     .\" HTML <a href="#conditions">
1333     .\" </a>
1334     conditions,
1335     .\"
1336     can be made by name as well as by number.
1337 nigel 75 .P
1338 nigel 91 Names consist of up to 32 alphanumeric characters and underscores. Named
1339 nigel 93 capturing parentheses are still allocated numbers as well as names, exactly as
1340     if the names were not present. The PCRE API provides function calls for
1341     extracting the name-to-number translation table from a compiled pattern. There
1342     is also a convenience function for extracting a captured substring by name.
1343 nigel 91 .P
1344     By default, a name must be unique within a pattern, but it is possible to relax
1345 ph10 457 this constraint by setting the PCRE_DUPNAMES option at compile time. (Duplicate
1346 ph10 461 names are also always permitted for subpatterns with the same number, set up as
1347 ph10 457 described in the previous section.) Duplicate names can be useful for patterns
1348     where only one instance of the named parentheses can match. Suppose you want to
1349     match the name of a weekday, either as a 3-letter abbreviation or as the full
1350     name, and in both cases you want to extract the abbreviation. This pattern
1351     (ignoring the line breaks) does the job:
1352 nigel 91 .sp
1353 nigel 93 (?<DN>Mon|Fri|Sun)(?:day)?|
1354     (?<DN>Tue)(?:sday)?|
1355     (?<DN>Wed)(?:nesday)?|
1356     (?<DN>Thu)(?:rsday)?|
1357     (?<DN>Sat)(?:urday)?
1358 nigel 91 .sp
1359     There are five capturing substrings, but only one is ever set after a match.
1360 ph10 182 (An alternative way of solving this problem is to use a "branch reset"
1361 ph10 175 subpattern, as described in the previous section.)
1362     .P
1363 nigel 91 The convenience function for extracting the data by name returns the substring
1364 nigel 93 for the first (and in this example, the only) subpattern of that name that
1365 ph10 461 matched. This saves searching to find which numbered subpattern it was.
1366 ph10 459 .P
1367 ph10 488 If you make a back reference to a non-unique named subpattern from elsewhere in
1368 ph10 459 the pattern, the one that corresponds to the first occurrence of the name is
1369     used. In the absence of duplicate numbers (see the previous section) this is
1370     the one with the lowest number. If you use a named reference in a condition
1371     test (see the
1372     .\"
1373     .\" HTML <a href="#conditions">
1374     .\" </a>
1375     section about conditions
1376     .\"
1377 ph10 461 below), either to check whether a subpattern has matched, or to check for
1378 ph10 459 recursion, all subpatterns with the same name are tested. If the condition is
1379     true for any one of them, the overall condition is true. This is the same
1380     behaviour as testing by number. For further details of the interfaces for
1381     handling named subpatterns, see the
1382 nigel 63 .\" HREF
1383 nigel 75 \fBpcreapi\fP
1384 nigel 63 .\"
1385     documentation.
1386 ph10 385 .P
1387     \fBWarning:\fP You cannot use different names to distinguish between two
1388 ph10 457 subpatterns with the same number because PCRE uses only the numbers when
1389     matching. For this reason, an error is given at compile time if different names
1390     are given to subpatterns with the same number. However, you can give the same
1391     name to subpatterns with the same number, even when PCRE_DUPNAMES is not set.
1392 nigel 75 .
1393     .
1394 nigel 63 .SH REPETITION
1395     .rs
1396     .sp
1397     Repetition is specified by quantifiers, which can follow any of the following
1398     items:
1399 nigel 75 .sp
1400 nigel 63 a literal data character
1401 nigel 93 the dot metacharacter
1402 nigel 75 the \eC escape sequence
1403     the \eX escape sequence (in UTF-8 mode with Unicode properties)
1404 nigel 93 the \eR escape sequence
1405 nigel 75 an escape such as \ed that matches a single character
1406 nigel 63 a character class
1407     a back reference (see next section)
1408     a parenthesized subpattern (unless it is an assertion)
1409 ph10 461 a recursive or "subroutine" call to a subpattern
1410 nigel 75 .sp
1411 nigel 63 The general repetition quantifier specifies a minimum and maximum number of
1412     permitted matches, by giving the two numbers in curly brackets (braces),
1413     separated by a comma. The numbers must be less than 65536, and the first must
1414     be less than or equal to the second. For example:
1415 nigel 75 .sp
1416 nigel 63 z{2,4}
1417 nigel 75 .sp
1418 nigel 63 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1419     character. If the second number is omitted, but the comma is present, there is
1420     no upper limit; if the second number and the comma are both omitted, the
1421     quantifier specifies an exact number of required matches. Thus
1422 nigel 75 .sp
1423 nigel 63 [aeiou]{3,}
1424 nigel 75 .sp
1425 nigel 63 matches at least 3 successive vowels, but may match many more, while
1426 nigel 75 .sp
1427     \ed{8}
1428     .sp
1429 nigel 63 matches exactly 8 digits. An opening curly bracket that appears in a position
1430     where a quantifier is not allowed, or one that does not match the syntax of a
1431     quantifier, is taken as a literal character. For example, {,6} is not a
1432     quantifier, but a literal string of four characters.
1433 nigel 75 .P
1434 nigel 63 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1435 nigel 75 bytes. Thus, for example, \ex{100}{2} matches two UTF-8 characters, each of
1436     which is represented by a two-byte sequence. Similarly, when Unicode property
1437     support is available, \eX{3} matches three Unicode extended sequences, each of
1438     which may be several bytes long (and they may be of different lengths).
1439     .P
1440 nigel 63 The quantifier {0} is permitted, causing the expression to behave as if the
1441 ph10 345 previous item and the quantifier were not present. This may be useful for
1442     subpatterns that are referenced as
1443 ph10 335 .\" HTML <a href="#subpatternsassubroutines">
1444     .\" </a>
1445     subroutines
1446     .\"
1447 ph10 345 from elsewhere in the pattern. Items other than subpatterns that have a {0}
1448 ph10 335 quantifier are omitted from the compiled pattern.
1449 nigel 75 .P
1450 nigel 93 For convenience, the three most common quantifiers have single-character
1451     abbreviations:
1452 nigel 75 .sp
1453 nigel 63 * is equivalent to {0,}
1454     + is equivalent to {1,}
1455     ? is equivalent to {0,1}
1456 nigel 75 .sp
1457 nigel 63 It is possible to construct infinite loops by following a subpattern that can
1458     match no characters with a quantifier that has no upper limit, for example:
1459 nigel 75 .sp
1460 nigel 63 (a?)*
1461 nigel 75 .sp
1462 nigel 63 Earlier versions of Perl and PCRE used to give an error at compile time for
1463     such patterns. However, because there are cases where this can be useful, such
1464     patterns are now accepted, but if any repetition of the subpattern does in fact
1465     match no characters, the loop is forcibly broken.
1466 nigel 75 .P
1467 nigel 63 By default, the quantifiers are "greedy", that is, they match as much as
1468     possible (up to the maximum number of permitted times), without causing the
1469     rest of the pattern to fail. The classic example of where this gives problems
1470 nigel 75 is in trying to match comments in C programs. These appear between /* and */
1471     and within the comment, individual * and / characters may appear. An attempt to
1472     match C comments by applying the pattern
1473     .sp
1474     /\e*.*\e*/
1475     .sp
1476 nigel 63 to the string
1477 nigel 75 .sp
1478     /* first comment */ not comment /* second comment */
1479     .sp
1480 nigel 63 fails, because it matches the entire string owing to the greediness of the .*
1481     item.
1482 nigel 75 .P
1483 nigel 63 However, if a quantifier is followed by a question mark, it ceases to be
1484     greedy, and instead matches the minimum number of times possible, so the
1485     pattern
1486 nigel 75 .sp
1487     /\e*.*?\e*/
1488     .sp
1489 nigel 63 does the right thing with the C comments. The meaning of the various
1490     quantifiers is not otherwise changed, just the preferred number of matches.
1491     Do not confuse this use of question mark with its use as a quantifier in its
1492     own right. Because it has two uses, it can sometimes appear doubled, as in
1493 nigel 75 .sp
1494     \ed??\ed
1495     .sp
1496 nigel 63 which matches one digit by preference, but can match two if that is the only
1497     way the rest of the pattern matches.
1498 nigel 75 .P
1499 nigel 93 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1500 nigel 63 the quantifiers are not greedy by default, but individual ones can be made
1501     greedy by following them with a question mark. In other words, it inverts the
1502     default behaviour.
1503 nigel 75 .P
1504 nigel 63 When a parenthesized subpattern is quantified with a minimum repeat count that
1505 nigel 75 is greater than 1 or with a limited maximum, more memory is required for the
1506 nigel 63 compiled pattern, in proportion to the size of the minimum or maximum.
1507 nigel 75 .P
1508 nigel 63 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1509 nigel 93 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1510 nigel 63 implicitly anchored, because whatever follows will be tried against every
1511     character position in the subject string, so there is no point in retrying the
1512     overall match at any position after the first. PCRE normally treats such a
1513 nigel 75 pattern as though it were preceded by \eA.
1514     .P
1515 nigel 63 In cases where it is known that the subject string contains no newlines, it is
1516     worth setting PCRE_DOTALL in order to obtain this optimization, or
1517     alternatively using ^ to indicate anchoring explicitly.
1518 nigel 75 .P
1519 nigel 63 However, there is one situation where the optimization cannot be used. When .*
1520 ph10 488 is inside capturing parentheses that are the subject of a back reference
1521 nigel 93 elsewhere in the pattern, a match at the start may fail where a later one
1522     succeeds. Consider, for example:
1523 nigel 75 .sp
1524     (.*)abc\e1
1525     .sp
1526 nigel 63 If the subject is "xyz123abc123" the match point is the fourth character. For
1527     this reason, such a pattern is not implicitly anchored.
1528 nigel 75 .P
1529 nigel 63 When a capturing subpattern is repeated, the value captured is the substring
1530     that matched the final iteration. For example, after
1531 nigel 75 .sp
1532     (tweedle[dume]{3}\es*)+
1533     .sp
1534 nigel 63 has matched "tweedledum tweedledee" the value of the captured substring is
1535     "tweedledee". However, if there are nested capturing subpatterns, the
1536     corresponding captured values may have been set in previous iterations. For
1537     example, after
1538 nigel 75 .sp
1539 nigel 63 /(a|(b))+/
1540 nigel 75 .sp
1541 nigel 63 matches "aba" the value of the second captured substring is "b".
1542 nigel 75 .
1543     .
1544     .\" HTML <a name="atomicgroup"></a>
1546 nigel 63 .rs
1547     .sp
1548 nigel 93 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1549     repetition, failure of what follows normally causes the repeated item to be
1550     re-evaluated to see if a different number of repeats allows the rest of the
1551     pattern to match. Sometimes it is useful to prevent this, either to change the
1552     nature of the match, or to cause it fail earlier than it otherwise might, when
1553     the author of the pattern knows there is no point in carrying on.
1554 nigel 75 .P
1555     Consider, for example, the pattern \ed+foo when applied to the subject line
1556     .sp
1557 nigel 63 123456bar
1558 nigel 75 .sp
1559 nigel 63 After matching all 6 digits and then failing to match "foo", the normal
1560 nigel 75 action of the matcher is to try again with only 5 digits matching the \ed+
1561 nigel 63 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1562     (a term taken from Jeffrey Friedl's book) provides the means for specifying
1563     that once a subpattern has matched, it is not to be re-evaluated in this way.
1564 nigel 75 .P
1565 nigel 93 If we use atomic grouping for the previous example, the matcher gives up
1566 nigel 63 immediately on failing to match "foo" the first time. The notation is a kind of
1567     special parenthesis, starting with (?> as in this example:
1568 nigel 75 .sp
1569     (?>\ed+)foo
1570     .sp
1571 nigel 63 This kind of parenthesis "locks up" the part of the pattern it contains once
1572     it has matched, and a failure further into the pattern is prevented from
1573     backtracking into it. Backtracking past it to previous items, however, works as
1574     normal.
1575 nigel 75 .P
1576 nigel 63 An alternative description is that a subpattern of this type matches the string
1577     of characters that an identical standalone pattern would match, if anchored at
1578     the current point in the subject string.
1579 nigel 75 .P
1580 nigel 63 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1581     the above example can be thought of as a maximizing repeat that must swallow
1582 nigel 75 everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1583 nigel 63 number of digits they match in order to make the rest of the pattern match,
1584 nigel 75 (?>\ed+) can only match an entire sequence of digits.
1585     .P
1586 nigel 63 Atomic groups in general can of course contain arbitrarily complicated
1587     subpatterns, and can be nested. However, when the subpattern for an atomic
1588     group is just a single repeated item, as in the example above, a simpler
1589     notation, called a "possessive quantifier" can be used. This consists of an
1590     additional + character following a quantifier. Using this notation, the
1591     previous example can be rewritten as
1592 nigel 75 .sp
1593     \ed++foo
1594     .sp
1595 ph10 208 Note that a possessive quantifier can be used with an entire group, for
1596     example:
1597     .sp
1598     (abc|xyz){2,3}+
1599     .sp
1600 nigel 63 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1601     option is ignored. They are a convenient notation for the simpler forms of
1602 nigel 93 atomic group. However, there is no difference in the meaning of a possessive
1603     quantifier and the equivalent atomic group, though there may be a performance
1604     difference; possessive quantifiers should be slightly faster.
1605 nigel 75 .P
1606 nigel 93 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1607     Jeffrey Friedl originated the idea (and the name) in the first edition of his
1608     book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1609     package, and PCRE copied it from there. It ultimately found its way into Perl
1610     at release 5.10.
1611 nigel 75 .P
1612 nigel 93 PCRE has an optimization that automatically "possessifies" certain simple
1613     pattern constructs. For example, the sequence A+B is treated as A++B because
1614     there is no point in backtracking into a sequence of A's when B must follow.
1615     .P
1616 nigel 63 When a pattern contains an unlimited repeat inside a subpattern that can itself
1617     be repeated an unlimited number of times, the use of an atomic group is the
1618     only way to avoid some failing matches taking a very long time indeed. The
1619     pattern
1620 nigel 75 .sp
1621     (\eD+|<\ed+>)*[!?]
1622     .sp
1623 nigel 63 matches an unlimited number of substrings that either consist of non-digits, or
1624     digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1625     quickly. However, if it is applied to
1626 nigel 75 .sp
1627 nigel 63 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1628 nigel 75 .sp
1629 nigel 63 it takes a long time before reporting failure. This is because the string can
1630 nigel 75 be divided between the internal \eD+ repeat and the external * repeat in a
1631     large number of ways, and all have to be tried. (The example uses [!?] rather
1632     than a single character at the end, because both PCRE and Perl have an
1633     optimization that allows for fast failure when a single character is used. They
1634     remember the last single character that is required for a match, and fail early
1635     if it is not present in the string.) If the pattern is changed so that it uses
1636     an atomic group, like this:
1637     .sp
1638     ((?>\eD+)|<\ed+>)*[!?]
1639     .sp
1640 nigel 63 sequences of non-digits cannot be broken, and failure happens quickly.
1641 nigel 75 .
1642     .
1643     .\" HTML <a name="backreferences"></a>
1645 nigel 63 .rs
1646     .sp
1647     Outside a character class, a backslash followed by a digit greater than 0 (and
1648     possibly further digits) is a back reference to a capturing subpattern earlier
1649     (that is, to its left) in the pattern, provided there have been that many
1650     previous capturing left parentheses.
1651 nigel 75 .P
1652 nigel 63 However, if the decimal number following the backslash is less than 10, it is
1653     always taken as a back reference, and causes an error only if there are not
1654     that many capturing left parentheses in the entire pattern. In other words, the
1655     parentheses that are referenced need not be to the left of the reference for
1656 nigel 91 numbers less than 10. A "forward back reference" of this type can make sense
1657     when a repetition is involved and the subpattern to the right has participated
1658     in an earlier iteration.
1659     .P
1660 nigel 93 It is not possible to have a numerical "forward back reference" to a subpattern
1661     whose number is 10 or more using this syntax because a sequence such as \e50 is
1662     interpreted as a character defined in octal. See the subsection entitled
1663 nigel 91 "Non-printing characters"
1664 nigel 75 .\" HTML <a href="#digitsafterbackslash">
1665     .\" </a>
1666     above
1667     .\"
1668 nigel 93 for further details of the handling of digits following a backslash. There is
1669     no such problem when named parentheses are used. A back reference to any
1670     subpattern is possible using named parentheses (see below).
1671 nigel 75 .P
1672 nigel 93 Another way of avoiding the ambiguity inherent in the use of digits following a
1673     backslash is to use the \eg escape sequence, which is a feature introduced in
1674 ph10 208 Perl 5.10. This escape must be followed by an unsigned number or a negative
1675     number, optionally enclosed in braces. These examples are all identical:
1676 nigel 93 .sp
1677     (ring), \e1
1678     (ring), \eg1
1679     (ring), \eg{1}
1680     .sp
1681 ph10 208 An unsigned number specifies an absolute reference without the ambiguity that
1682     is present in the older syntax. It is also useful when literal digits follow
1683     the reference. A negative number is a relative reference. Consider this
1684     example:
1685 nigel 93 .sp
1686     (abc(def)ghi)\eg{-1}
1687     .sp
1688     The sequence \eg{-1} is a reference to the most recently started capturing
1689     subpattern before \eg, that is, is it equivalent to \e2. Similarly, \eg{-2}
1690     would be equivalent to \e1. The use of relative references can be helpful in
1691     long patterns, and also in patterns that are created by joining together
1692     fragments that contain references within themselves.
1693     .P
1694 nigel 63 A back reference matches whatever actually matched the capturing subpattern in
1695     the current subject string, rather than anything matching the subpattern
1696     itself (see
1697     .\" HTML <a href="#subpatternsassubroutines">
1698     .\" </a>
1699     "Subpatterns as subroutines"
1700     .\"
1701     below for a way of doing that). So the pattern
1702 nigel 75 .sp
1703     (sens|respons)e and \e1ibility
1704     .sp
1705 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1706     "sense and responsibility". If caseful matching is in force at the time of the
1707     back reference, the case of letters is relevant. For example,
1708 nigel 75 .sp
1709     ((?i)rah)\es+\e1
1710     .sp
1711 nigel 63 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1712     capturing subpattern is matched caselessly.
1713 nigel 75 .P
1714 ph10 171 There are several different ways of writing back references to named
1715     subpatterns. The .NET syntax \ek{name} and the Perl syntax \ek<name> or
1716     \ek'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1717     back reference syntax, in which \eg can be used for both numeric and named
1718     references, is also supported. We could rewrite the above example in any of
1719 nigel 93 the following ways:
1720 nigel 75 .sp
1721 nigel 93 (?<p1>(?i)rah)\es+\ek<p1>
1722 ph10 171 (?'p1'(?i)rah)\es+\ek{p1}
1723 nigel 91 (?P<p1>(?i)rah)\es+(?P=p1)
1724 ph10 171 (?<p1>(?i)rah)\es+\eg{p1}
1725 nigel 75 .sp
1726 nigel 91 A subpattern that is referenced by name may appear in the pattern before or
1727     after the reference.
1728     .P
1729 nigel 63 There may be more than one back reference to the same subpattern. If a
1730     subpattern has not actually been used in a particular match, any back
1731 ph10 456 references to it always fail by default. For example, the pattern
1732 nigel 75 .sp
1733     (a|(bc))\e2
1734     .sp
1735 ph10 461 always fails if it starts to match "a" rather than "bc". However, if the
1736     PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back reference to an
1737 ph10 456 unset value matches an empty string.
1738     .P
1739     Because there may be many capturing parentheses in a pattern, all digits
1740     following a backslash are taken as part of a potential back reference number.
1741     If the pattern continues with a digit character, some delimiter must be used to
1742     terminate the back reference. If the PCRE_EXTENDED option is set, this can be
1743     whitespace. Otherwise, the \eg{ syntax or an empty comment (see
1744 nigel 75 .\" HTML <a href="#comments">
1745     .\" </a>
1746     "Comments"
1747     .\"
1748     below) can be used.
1749 ph10 488 .
1750     .SS "Recursive back references"
1751     .rs
1752     .sp
1753 nigel 63 A back reference that occurs inside the parentheses to which it refers fails
1754 nigel 75 when the subpattern is first used, so, for example, (a\e1) never matches.
1755 nigel 63 However, such references can be useful inside repeated subpatterns. For
1756     example, the pattern
1757 nigel 75 .sp
1758     (a|b\e1)+
1759     .sp
1760 nigel 63 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1761     the subpattern, the back reference matches the character string corresponding
1762     to the previous iteration. In order for this to work, the pattern must be such
1763     that the first iteration does not need to match the back reference. This can be
1764     done using alternation, as in the example above, or by a quantifier with a
1765     minimum of zero.
1766 ph10 488 .P
1767     Back references of this type cause the group that they reference to be treated
1768     as an
1769     .\" HTML <a href="#atomicgroup">
1770     .\" </a>
1771     atomic group.
1772     .\"
1773     Once the whole group has been matched, a subsequent matching failure cannot
1774     cause backtracking into the middle of the group.
1775 nigel 75 .
1776     .
1777     .\" HTML <a name="bigassertions"></a>
1778 nigel 63 .SH ASSERTIONS
1779     .rs
1780     .sp
1781     An assertion is a test on the characters following or preceding the current
1782     matching point that does not actually consume any characters. The simple
1783 nigel 75 assertions coded as \eb, \eB, \eA, \eG, \eZ, \ez, ^ and $ are described
1784     .\" HTML <a href="#smallassertions">
1785     .\" </a>
1786     above.
1787     .\"
1788     .P
1789 nigel 63 More complicated assertions are coded as subpatterns. There are two kinds:
1790     those that look ahead of the current position in the subject string, and those
1791 nigel 75 that look behind it. An assertion subpattern is matched in the normal way,
1792     except that it does not cause the current matching position to be changed.
1793     .P
1794     Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1795     because it makes no sense to assert the same thing several times. If any kind
1796     of assertion contains capturing subpatterns within it, these are counted for
1797     the purposes of numbering the capturing subpatterns in the whole pattern.
1798     However, substring capturing is carried out only for positive assertions,
1799     because it does not make sense for negative assertions.
1800     .
1801     .
1802     .SS "Lookahead assertions"
1803     .rs
1804     .sp
1805 nigel 91 Lookahead assertions start with (?= for positive assertions and (?! for
1806     negative assertions. For example,
1807 nigel 75 .sp
1808     \ew+(?=;)
1809     .sp
1810 nigel 63 matches a word followed by a semicolon, but does not include the semicolon in
1811     the match, and
1812 nigel 75 .sp
1813 nigel 63 foo(?!bar)
1814 nigel 75 .sp
1815 nigel 63 matches any occurrence of "foo" that is not followed by "bar". Note that the
1816     apparently similar pattern
1817 nigel 75 .sp
1818 nigel 63 (?!foo)bar
1819 nigel 75 .sp
1820 nigel 63 does not find an occurrence of "bar" that is preceded by something other than
1821     "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1822     (?!foo) is always true when the next three characters are "bar". A
1823 nigel 75 lookbehind assertion is needed to achieve the other effect.
1824     .P
1825 nigel 63 If you want to force a matching failure at some point in a pattern, the most
1826     convenient way to do it is with (?!) because an empty string always matches, so
1827     an assertion that requires there not to be an empty string must always fail.
1828 ph10 456 The Perl 5.10 backtracking control verb (*FAIL) or (*F) is essentially a
1829     synonym for (?!).
1830 nigel 75 .
1831     .
1832     .\" HTML <a name="lookbehind"></a>
1833     .SS "Lookbehind assertions"
1834     .rs
1835     .sp
1836 nigel 63 Lookbehind assertions start with (?<= for positive assertions and (?<! for
1837     negative assertions. For example,
1838 nigel 75 .sp
1839 nigel 63 (?<!foo)bar
1840 nigel 75 .sp
1841 nigel 63 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1842     a lookbehind assertion are restricted such that all the strings it matches must
1843 nigel 91 have a fixed length. However, if there are several top-level alternatives, they
1844     do not all have to have the same fixed length. Thus
1845 nigel 75 .sp
1846 nigel 63 (?<=bullock|donkey)
1847 nigel 75 .sp
1848 nigel 63 is permitted, but
1849 nigel 75 .sp
1850 nigel 63 (?<!dogs?|cats?)
1851 nigel 75 .sp
1852 nigel 63 causes an error at compile time. Branches that match different length strings
1853     are permitted only at the top level of a lookbehind assertion. This is an
1854 ph10 454 extension compared with Perl (5.8 and 5.10), which requires all branches to
1855 nigel 63 match the same length of string. An assertion such as
1856 nigel 75 .sp
1857 nigel 63 (?<=ab(c|de))
1858 nigel 75 .sp
1859 nigel 63 is not permitted, because its single top-level branch can match two different
1860 ph10 454 lengths, but it is acceptable to PCRE if rewritten to use two top-level
1861     branches:
1862 nigel 75 .sp
1863 nigel 63 (?<=abc|abde)
1864 nigel 75 .sp
1865 ph10 172 In some cases, the Perl 5.10 escape sequence \eK
1866 ph10 168 .\" HTML <a href="#resetmatchstart">
1867     .\" </a>
1868     (see above)
1869     .\"
1870 ph10 461 can be used instead of a lookbehind assertion to get round the fixed-length
1871 ph10 454 restriction.
1872 ph10 168 .P
1873 nigel 63 The implementation of lookbehind assertions is, for each alternative, to
1874 nigel 93 temporarily move the current position back by the fixed length and then try to
1875 nigel 63 match. If there are insufficient characters before the current position, the
1876 nigel 93 assertion fails.
1877 nigel 75 .P
1878     PCRE does not allow the \eC escape (which matches a single byte in UTF-8 mode)
1879 nigel 63 to appear in lookbehind assertions, because it makes it impossible to calculate
1880 nigel 93 the length of the lookbehind. The \eX and \eR escapes, which can match
1881     different numbers of bytes, are also not permitted.
1882 nigel 75 .P
1883 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
1884     .\" </a>
1885     "Subroutine"
1886     .\"
1887     calls (see below) such as (?2) or (?&X) are permitted in lookbehinds, as long
1888 ph10 461 as the subpattern matches a fixed-length string.
1889 ph10 454 .\" HTML <a href="#recursion">
1890     .\" </a>
1891     Recursion,
1892     .\"
1893     however, is not supported.
1894     .P
1895 nigel 93 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1896 ph10 456 specify efficient matching of fixed-length strings at the end of subject
1897     strings. Consider a simple pattern such as
1898 nigel 75 .sp
1899 nigel 63 abcd$
1900 nigel 75 .sp
1901 nigel 63 when applied to a long string that does not match. Because matching proceeds
1902     from left to right, PCRE will look for each "a" in the subject and then see if
1903     what follows matches the rest of the pattern. If the pattern is specified as
1904 nigel 75 .sp
1905 nigel 63 ^.*abcd$
1906 nigel 75 .sp
1907 nigel 63 the initial .* matches the entire string at first, but when this fails (because
1908     there is no following "a"), it backtracks to match all but the last character,
1909     then all but the last two characters, and so on. Once again the search for "a"
1910     covers the entire string, from right to left, so we are no better off. However,
1911     if the pattern is written as
1912 nigel 75 .sp
1913 nigel 63 ^.*+(?<=abcd)
1914 nigel 75 .sp
1915 nigel 93 there can be no backtracking for the .*+ item; it can match only the entire
1916 nigel 63 string. The subsequent lookbehind assertion does a single test on the last four
1917     characters. If it fails, the match fails immediately. For long strings, this
1918     approach makes a significant difference to the processing time.
1919 nigel 75 .
1920     .
1921     .SS "Using multiple assertions"
1922     .rs
1923     .sp
1924 nigel 63 Several assertions (of any sort) may occur in succession. For example,
1925 nigel 75 .sp
1926     (?<=\ed{3})(?<!999)foo
1927     .sp
1928 nigel 63 matches "foo" preceded by three digits that are not "999". Notice that each of
1929     the assertions is applied independently at the same point in the subject
1930     string. First there is a check that the previous three characters are all
1931     digits, and then there is a check that the same three characters are not "999".
1932 nigel 75 This pattern does \fInot\fP match "foo" preceded by six characters, the first
1933 nigel 63 of which are digits and the last three of which are not "999". For example, it
1934     doesn't match "123abcfoo". A pattern to do that is
1935 nigel 75 .sp
1936     (?<=\ed{3}...)(?<!999)foo
1937     .sp
1938 nigel 63 This time the first assertion looks at the preceding six characters, checking
1939     that the first three are digits, and then the second assertion checks that the
1940     preceding three characters are not "999".
1941 nigel 75 .P
1942 nigel 63 Assertions can be nested in any combination. For example,
1943 nigel 75 .sp
1944 nigel 63 (?<=(?<!foo)bar)baz
1945 nigel 75 .sp
1946 nigel 63 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1947     preceded by "foo", while
1948 nigel 75 .sp
1949     (?<=\ed{3}(?!999)...)foo
1950     .sp
1951     is another pattern that matches "foo" preceded by three digits and any three
1952 nigel 63 characters that are not "999".
1953 nigel 75 .
1954     .
1955 nigel 91 .\" HTML <a name="conditions"></a>
1957 nigel 63 .rs
1958     .sp
1959     It is possible to cause the matching process to obey a subpattern
1960     conditionally or to choose between two alternative subpatterns, depending on
1961 ph10 461 the result of an assertion, or whether a specific capturing subpattern has
1962 ph10 456 already been matched. The two possible forms of conditional subpattern are:
1963 nigel 75 .sp
1964 nigel 63 (?(condition)yes-pattern)
1965     (?(condition)yes-pattern|no-pattern)
1966 nigel 75 .sp
1967 nigel 63 If the condition is satisfied, the yes-pattern is used; otherwise the
1968     no-pattern (if present) is used. If there are more than two alternatives in the
1969     subpattern, a compile-time error occurs.
1970 nigel 75 .P
1971 nigel 93 There are four kinds of condition: references to subpatterns, references to
1972     recursion, a pseudo-condition called DEFINE, and assertions.
1973     .
1974     .SS "Checking for a used subpattern by number"
1975     .rs
1976     .sp
1977     If the text between the parentheses consists of a sequence of digits, the
1978 ph10 456 condition is true if a capturing subpattern of that number has previously
1979 ph10 461 matched. If there is more than one capturing subpattern with the same number
1980     (see the earlier
1981 ph10 456 .\"
1982     .\" HTML <a href="#recursion">
1983     .\" </a>
1984     section about duplicate subpattern numbers),
1985     .\"
1986     the condition is true if any of them have been set. An alternative notation is
1987     to precede the digits with a plus or minus sign. In this case, the subpattern
1988     number is relative rather than absolute. The most recently opened parentheses
1989     can be referenced by (?(-1), the next most recent by (?(-2), and so on. In
1990     looping constructs it can also make sense to refer to subsequent groups with
1991     constructs such as (?(+2).
1992 nigel 91 .P
1993     Consider the following pattern, which contains non-significant white space to
1994     make it more readable (assume the PCRE_EXTENDED option) and to divide it into
1995     three parts for ease of discussion:
1996 nigel 75 .sp
1997     ( \e( )? [^()]+ (?(1) \e) )
1998     .sp
1999 nigel 63 The first part matches an optional opening parenthesis, and if that
2000     character is present, sets it as the first captured substring. The second part
2001     matches one or more characters that are not parentheses. The third part is a
2002     conditional subpattern that tests whether the first set of parentheses matched
2003     or not. If they did, that is, if subject started with an opening parenthesis,
2004     the condition is true, and so the yes-pattern is executed and a closing
2005     parenthesis is required. Otherwise, since no-pattern is not present, the
2006     subpattern matches nothing. In other words, this pattern matches a sequence of
2007 nigel 93 non-parentheses, optionally enclosed in parentheses.
2008 ph10 167 .P
2009 ph10 172 If you were embedding this pattern in a larger one, you could use a relative
2010 ph10 167 reference:
2011     .sp
2012     ...other stuff... ( \e( )? [^()]+ (?(-1) \e) ) ...
2013     .sp
2014     This makes the fragment independent of the parentheses in the larger pattern.
2015 nigel 93 .
2016     .SS "Checking for a used subpattern by name"
2017     .rs
2018 nigel 91 .sp
2019 nigel 93 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a used
2020     subpattern by name. For compatibility with earlier versions of PCRE, which had
2021     this facility before Perl, the syntax (?(name)...) is also recognized. However,
2022     there is a possible ambiguity with this syntax, because subpattern names may
2023     consist entirely of digits. PCRE looks first for a named subpattern; if it
2024     cannot find one and the name consists entirely of digits, PCRE looks for a
2025     subpattern of that number, which must be greater than zero. Using subpattern
2026     names that consist entirely of digits is not recommended.
2027     .P
2028     Rewriting the above example to use a named subpattern gives this:
2029 nigel 91 .sp
2030 nigel 93 (?<OPEN> \e( )? [^()]+ (?(<OPEN>) \e) )
2031     .sp
2032 ph10 461 If the name used in a condition of this kind is a duplicate, the test is
2033     applied to all subpatterns of the same name, and is true if any one of them has
2034 ph10 459 matched.
2035 nigel 93 .
2036     .SS "Checking for pattern recursion"
2037     .rs
2038     .sp
2039 nigel 91 If the condition is the string (R), and there is no subpattern with the name R,
2040 nigel 93 the condition is true if a recursive call to the whole pattern or any
2041     subpattern has been made. If digits or a name preceded by ampersand follow the
2042     letter R, for example:
2043     .sp
2044     (?(R3)...) or (?(R&name)...)
2045     .sp
2046 ph10 456 the condition is true if the most recent recursion is into a subpattern whose
2047 nigel 93 number or name is given. This condition does not check the entire recursion
2048 ph10 461 stack. If the name used in a condition of this kind is a duplicate, the test is
2049     applied to all subpatterns of the same name, and is true if any one of them is
2050     the most recent recursion.
2051 nigel 75 .P
2052 ph10 461 At "top level", all these recursion test conditions are false.
2053 ph10 454 .\" HTML <a href="#recursion">
2054     .\" </a>
2055 ph10 459 The syntax for recursive patterns
2056 ph10 454 .\"
2057 ph10 459 is described below.
2058 nigel 93 .
2059     .SS "Defining subpatterns for use by reference only"
2060     .rs
2061     .sp
2062     If the condition is the string (DEFINE), and there is no subpattern with the
2063     name DEFINE, the condition is always false. In this case, there may be only one
2064     alternative in the subpattern. It is always skipped if control reaches this
2065     point in the pattern; the idea of DEFINE is that it can be used to define
2066 ph10 461 "subroutines" that can be referenced from elsewhere. (The use of
2067 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
2068     .\" </a>
2069     "subroutines"
2070     .\"
2071 nigel 93 is described below.) For example, a pattern to match an IPv4 address could be
2072     written like this (ignore whitespace and line breaks):
2073     .sp
2074     (?(DEFINE) (?<byte> 2[0-4]\ed | 25[0-5] | 1\ed\ed | [1-9]?\ed) )
2075     \eb (?&byte) (\e.(?&byte)){3} \eb
2076     .sp
2077     The first part of the pattern is a DEFINE group inside which a another group
2078     named "byte" is defined. This matches an individual component of an IPv4
2079     address (a number less than 256). When matching takes place, this part of the
2080 ph10 456 pattern is skipped because DEFINE acts like a false condition. The rest of the
2081     pattern uses references to the named group to match the four dot-separated
2082     components of an IPv4 address, insisting on a word boundary at each end.
2083 nigel 93 .
2084     .SS "Assertion conditions"
2085     .rs
2086     .sp
2087     If the condition is not in any of the above formats, it must be an assertion.
2088 nigel 63 This may be a positive or negative lookahead or lookbehind assertion. Consider
2089     this pattern, again containing non-significant white space, and with the two
2090     alternatives on the second line:
2091 nigel 75 .sp
2092 nigel 63 (?(?=[^a-z]*[a-z])
2093 nigel 75 \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
2094     .sp
2095 nigel 63 The condition is a positive lookahead assertion that matches an optional
2096     sequence of non-letters followed by a letter. In other words, it tests for the
2097     presence of at least one letter in the subject. If a letter is found, the
2098     subject is matched against the first alternative; otherwise it is matched
2099     against the second. This pattern matches strings in one of the two forms
2100     dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
2101 nigel 75 .
2102     .
2103     .\" HTML <a name="comments"></a>
2104 nigel 63 .SH COMMENTS
2105     .rs
2106     .sp
2107 nigel 75 The sequence (?# marks the start of a comment that continues up to the next
2108 nigel 63 closing parenthesis. Nested parentheses are not permitted. The characters
2109     that make up a comment play no part in the pattern matching at all.
2110 nigel 75 .P
2111 nigel 63 If the PCRE_EXTENDED option is set, an unescaped # character outside a
2112 nigel 91 character class introduces a comment that continues to immediately after the
2113 ph10 556 next newline character or character sequence in the pattern. Which characters
2114     are interpreted as newlines is controlled by the options passed to
2115     \fBpcre_compile()\fP or by a special sequence at the start of the pattern, as
2116     described in the section entitled
2117     .\" HTML <a href="#recursion">
2118     .\" </a>
2119     "Newline conventions"
2120     .\"
2121     above. Note that end of a comment is a literal newline sequence in the pattern;
2122     escape sequences that happen to represent a newline do not terminate a comment.
2123     For example, consider this pattern when PCRE_EXTENDED is set, and the default
2124     newline convention is in force:
2125     .sp
2126     abc #comment \en still comment
2127     .sp
2128     On encountering the # character, \fBpcre_compile()\fP skips along, looking for
2129     a newline in the pattern. The sequence \en is still literal at this stage, so
2130     it does not terminate the comment. Only an actual character with the code value
2131     0x0a does so.
2132 nigel 75 .
2133     .
2134 nigel 91 .\" HTML <a name="recursion"></a>
2135 nigel 75 .SH "RECURSIVE PATTERNS"
2136 nigel 63 .rs
2137     .sp
2138     Consider the problem of matching a string in parentheses, allowing for
2139     unlimited nested parentheses. Without the use of recursion, the best that can
2140     be done is to use a pattern that matches up to some fixed depth of nesting. It
2141 nigel 93 is not possible to handle an arbitrary nesting depth.
2142     .P
2143     For some time, Perl has provided a facility that allows regular expressions to
2144     recurse (amongst other things). It does this by interpolating Perl code in the
2145     expression at run time, and the code can refer to the expression itself. A Perl
2146     pattern using code interpolation to solve the parentheses problem can be
2147     created like this:
2148 nigel 75 .sp
2149     $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
2150     .sp
2151 nigel 63 The (?p{...}) item interpolates Perl code at run time, and in this case refers
2152 nigel 93 recursively to the pattern in which it appears.
2153 nigel 75 .P
2154 nigel 93 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
2155     supports special syntax for recursion of the entire pattern, and also for
2156     individual subpattern recursion. After its introduction in PCRE and Python,
2157 ph10 453 this kind of recursion was subsequently introduced into Perl at release 5.10.
2158 nigel 75 .P
2159 nigel 93 A special item that consists of (? followed by a number greater than zero and a
2160     closing parenthesis is a recursive call of the subpattern of the given number,
2161 ph10 461 provided that it occurs inside that subpattern. (If not, it is a
2162 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
2163     .\" </a>
2164     "subroutine"
2165     .\"
2166 nigel 93 call, which is described in the next section.) The special item (?R) or (?0) is
2167     a recursive call of the entire regular expression.
2168 nigel 87 .P
2169     This PCRE pattern solves the nested parentheses problem (assume the
2170     PCRE_EXTENDED option is set so that white space is ignored):
2171 nigel 75 .sp
2172 ph10 456 \e( ( [^()]++ | (?R) )* \e)
2173 nigel 75 .sp
2174 nigel 63 First it matches an opening parenthesis. Then it matches any number of
2175     substrings which can either be a sequence of non-parentheses, or a recursive
2176 nigel 87 match of the pattern itself (that is, a correctly parenthesized substring).
2177 ph10 461 Finally there is a closing parenthesis. Note the use of a possessive quantifier
2178 ph10 456 to avoid backtracking into sequences of non-parentheses.
2179 nigel 75 .P
2180 nigel 63 If this were part of a larger pattern, you would not want to recurse the entire
2181     pattern, so instead you could use this:
2182 nigel 75 .sp
2183 ph10 456 ( \e( ( [^()]++ | (?1) )* \e) )
2184 nigel 75 .sp
2185 nigel 63 We have put the pattern into parentheses, and caused the recursion to refer to
2186 ph10 172 them instead of the whole pattern.
2187 ph10 166 .P
2188     In a larger pattern, keeping track of parenthesis numbers can be tricky. This
2189 ph10 456 is made easier by the use of relative references (a Perl 5.10 feature).
2190 ph10 166 Instead of (?1) in the pattern above you can write (?-2) to refer to the second
2191     most recently opened parentheses preceding the recursion. In other words, a
2192     negative number counts capturing parentheses leftwards from the point at which
2193     it is encountered.
2194     .P
2195     It is also possible to refer to subsequently opened parentheses, by writing
2196     references such as (?+2). However, these cannot be recursive because the
2197     reference is not inside the parentheses that are referenced. They are always
2198 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
2199     .\" </a>
2200     "subroutine"
2201     .\"
2202     calls, as described in the next section.
2203 ph10 166 .P
2204     An alternative approach is to use named parentheses instead. The Perl syntax
2205     for this is (?&name); PCRE's earlier syntax (?P>name) is also supported. We
2206     could rewrite the above example as follows:
2207 nigel 75 .sp
2208 ph10 456 (?<pn> \e( ( [^()]++ | (?&pn) )* \e) )
2209 nigel 75 .sp
2210 nigel 93 If there is more than one subpattern with the same name, the earliest one is
2211 ph10 172 used.
2212 ph10 166 .P
2213     This particular example pattern that we have been looking at contains nested
2214 ph10 456 unlimited repeats, and so the use of a possessive quantifier for matching
2215     strings of non-parentheses is important when applying the pattern to strings
2216     that do not match. For example, when this pattern is applied to
2217 nigel 75 .sp
2218 nigel 63 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
2219 nigel 75 .sp
2220 ph10 456 it yields "no match" quickly. However, if a possessive quantifier is not used,
2221 nigel 63 the match runs for a very long time indeed because there are so many different
2222     ways the + and * repeats can carve up the subject, and all have to be tested
2223     before failure can be reported.
2224 nigel 75 .P
2225 ph10 464 At the end of a match, the values of capturing parentheses are those from
2226     the outermost level. If you want to obtain intermediate values, a callout
2227     function can be used (see below and the
2228 nigel 63 .\" HREF
2229 nigel 75 \fBpcrecallout\fP
2230 nigel 63 .\"
2231     documentation). If the pattern above is matched against
2232 nigel 75 .sp
2233 nigel 63 (ab(cd)ef)
2234 nigel 75 .sp
2235 ph10 464 the value for the inner capturing parentheses (numbered 2) is "ef", which is
2236     the last value taken on at the top level. If a capturing subpattern is not
2237     matched at the top level, its final value is unset, even if it is (temporarily)
2238     set at a deeper level.
2239 nigel 75 .P
2240 ph10 464 If there are more than 15 capturing parentheses in a pattern, PCRE has to
2241     obtain extra memory to store data during a recursion, which it does by using
2242     \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no memory can
2243     be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
2244     .P
2245 nigel 63 Do not confuse the (?R) item with the condition (R), which tests for recursion.
2246     Consider this pattern, which matches text in angle brackets, allowing for
2247     arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2248     recursing), whereas any characters are permitted at the outer level.
2249 nigel 75 .sp
2250     < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
2251     .sp
2252 nigel 63 In this pattern, (?(R) is the start of a conditional subpattern, with two
2253     different alternatives for the recursive and non-recursive cases. The (?R) item
2254     is the actual recursive call.
2255 nigel 75 .
2256     .
2257 ph10 453 .\" HTML <a name="recursiondifference"></a>
2258     .SS "Recursion difference from Perl"
2259     .rs
2260     .sp
2261     In PCRE (like Python, but unlike Perl), a recursive subpattern call is always
2262     treated as an atomic group. That is, once it has matched some of the subject
2263     string, it is never re-entered, even if it contains untried alternatives and
2264 ph10 461 there is a subsequent matching failure. This can be illustrated by the
2265     following pattern, which purports to match a palindromic string that contains
2266 ph10 453 an odd number of characters (for example, "a", "aba", "abcba", "abcdcba"):
2267     .sp
2268     ^(.|(.)(?1)\e2)$
2269     .sp
2270 ph10 461 The idea is that it either matches a single character, or two identical
2271     characters surrounding a sub-palindrome. In Perl, this pattern works; in PCRE
2272 ph10 453 it does not if the pattern is longer than three characters. Consider the
2273     subject string "abcba":
2274     .P
2275 ph10 461 At the top level, the first character is matched, but as it is not at the end
2276 ph10 453 of the string, the first alternative fails; the second alternative is taken
2277     and the recursion kicks in. The recursive call to subpattern 1 successfully
2278     matches the next character ("b"). (Note that the beginning and end of line
2279     tests are not part of the recursion).
2280     .P
2281     Back at the top level, the next character ("c") is compared with what
2282 ph10 461 subpattern 2 matched, which was "a". This fails. Because the recursion is
2283 ph10 453 treated as an atomic group, there are now no backtracking points, and so the
2284     entire match fails. (Perl is able, at this point, to re-enter the recursion and
2285     try the second alternative.) However, if the pattern is written with the
2286     alternatives in the other order, things are different:
2287     .sp
2288     ^((.)(?1)\e2|.)$
2289     .sp
2290 ph10 461 This time, the recursing alternative is tried first, and continues to recurse
2291     until it runs out of characters, at which point the recursion fails. But this
2292     time we do have another alternative to try at the higher level. That is the big
2293 ph10 453 difference: in the previous case the remaining alternative is at a deeper
2294     recursion level, which PCRE cannot use.
2295     .P
2296 ph10 461 To change the pattern so that matches all palindromic strings, not just those
2297 ph10 453 with an odd number of characters, it is tempting to change the pattern to this:
2298     .sp
2299     ^((.)(?1)\e2|.?)$
2300     .sp
2301 ph10 461 Again, this works in Perl, but not in PCRE, and for the same reason. When a
2302     deeper recursion has matched a single character, it cannot be entered again in
2303     order to match an empty string. The solution is to separate the two cases, and
2304 ph10 453 write out the odd and even cases as alternatives at the higher level:
2305     .sp
2306     ^(?:((.)(?1)\e2|)|((.)(?3)\e4|.))
2307 ph10 461 .sp
2308     If you want to match typical palindromic phrases, the pattern has to ignore all
2309 ph10 453 non-word characters, which can be done like this:
2310     .sp
2311 ph10 461 ^\eW*+(?:((.)\eW*+(?1)\eW*+\e2|)|((.)\eW*+(?3)\eW*+\e4|\eW*+.\eW*+))\eW*+$
2312 ph10 453 .sp
2313 ph10 461 If run with the PCRE_CASELESS option, this pattern matches phrases such as "A
2314     man, a plan, a canal: Panama!" and it works well in both PCRE and Perl. Note
2315     the use of the possessive quantifier *+ to avoid backtracking into sequences of
2316 ph10 453 non-word characters. Without this, PCRE takes a great deal longer (ten times or
2317     more) to match typical phrases, and Perl takes so long that you think it has
2318     gone into a loop.
2319 ph10 456 .P
2320     \fBWARNING\fP: The palindrome-matching patterns above work only if the subject
2321     string does not start with a palindrome that is shorter than the entire string.
2322     For example, although "abcba" is correctly matched, if the subject is "ababa",
2323     PCRE finds the palindrome "aba" at the start, then fails at top level because
2324     the end of the string does not follow. Once again, it cannot jump back into the
2325     recursion to try other alternatives, so the entire match fails.
2326 ph10 453 .
2327     .
2328 nigel 63 .\" HTML <a name="subpatternsassubroutines"></a>
2330 nigel 63 .rs
2331     .sp
2332     If the syntax for a recursive subpattern reference (either by number or by
2333     name) is used outside the parentheses to which it refers, it operates like a
2334 nigel 93 subroutine in a programming language. The "called" subpattern may be defined
2335 ph10 166 before or after the reference. A numbered reference can be absolute or
2336     relative, as in these examples:
2337 nigel 75 .sp
2338 ph10 166 (...(absolute)...)...(?2)...
2339     (...(relative)...)...(?-1)...
2340 ph10 172 (...(?+1)...(relative)...
2341 ph10 166 .sp
2342     An earlier example pointed out that the pattern
2343     .sp
2344 nigel 75 (sens|respons)e and \e1ibility
2345     .sp
2346 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
2347     "sense and responsibility". If instead the pattern
2348 nigel 75 .sp
2349 nigel 63 (sens|respons)e and (?1)ibility
2350 nigel 75 .sp
2351 nigel 63 is used, it does match "sense and responsibility" as well as the other two
2352 nigel 93 strings. Another example is given in the discussion of DEFINE above.
2353 nigel 87 .P
2354 ph10 464 Like recursive subpatterns, a subroutine call is always treated as an atomic
2355 nigel 87 group. That is, once it has matched some of the subject string, it is never
2356     re-entered, even if it contains untried alternatives and there is a subsequent
2357 ph10 469 matching failure. Any capturing parentheses that are set during the subroutine
2358 ph10 464 call revert to their previous values afterwards.
2359 nigel 93 .P
2360     When a subpattern is used as a subroutine, processing options such as
2361     case-independence are fixed when the subpattern is defined. They cannot be
2362     changed for different calls. For example, consider this pattern:
2363     .sp
2364 ph10 166 (abc)(?i:(?-1))
2365 nigel 93 .sp
2366     It matches "abcabc". It does not match "abcABC" because the change of
2367     processing option does not affect the called subpattern.
2368 nigel 75 .
2369     .
2370 ph10 333 .\" HTML <a name="onigurumasubroutines"></a>
2372     .rs
2373     .sp
2374 ph10 345 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
2375     a number enclosed either in angle brackets or single quotes, is an alternative
2376     syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2377 ph10 333 are two of the examples used above, rewritten using this syntax:
2378     .sp
2379     (?<pn> \e( ( (?>[^()]+) | \eg<pn> )* \e) )
2380     (sens|respons)e and \eg'1'ibility
2381     .sp
2382 ph10 345 PCRE supports an extension to Oniguruma: if a number is preceded by a
2383 ph10 333 plus or a minus sign it is taken as a relative reference. For example:
2384     .sp
2385     (abc)(?i:\eg<-1>)
2386     .sp
2387 ph10 345 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
2388 ph10 333 synonymous. The former is a back reference; the latter is a subroutine call.
2389     .
2390     .
2391 nigel 63 .SH CALLOUTS
2392     .rs
2393     .sp
2394     Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2395     code to be obeyed in the middle of matching a regular expression. This makes it
2396     possible, amongst other things, to extract different substrings that match the
2397     same pair of parentheses when there is a repetition.
2398 nigel 75 .P
2399 nigel 63 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2400     code. The feature is called "callout". The caller of PCRE provides an external
2401 nigel 75 function by putting its entry point in the global variable \fIpcre_callout\fP.
2402 nigel 63 By default, this variable contains NULL, which disables all calling out.
2403 nigel 75 .P
2404 nigel 63 Within a regular expression, (?C) indicates the points at which the external
2405     function is to be called. If you want to identify different callout points, you
2406     can put a number less than 256 after the letter C. The default value is zero.
2407     For example, this pattern has two callout points:
2408 nigel 75 .sp
2409 ph10 155 (?C1)abc(?C2)def
2410 nigel 75 .sp
2411     If the PCRE_AUTO_CALLOUT flag is passed to \fBpcre_compile()\fP, callouts are
2412     automatically installed before each item in the pattern. They are all numbered
2413     255.
2414     .P
2415     During matching, when PCRE reaches a callout point (and \fIpcre_callout\fP is
2416 nigel 63 set), the external function is called. It is provided with the number of the
2417 nigel 75 callout, the position in the pattern, and, optionally, one item of data
2418     originally supplied by the caller of \fBpcre_exec()\fP. The callout function
2419     may cause matching to proceed, to backtrack, or to fail altogether. A complete
2420     description of the interface to the callout function is given in the
2421 nigel 63 .\" HREF
2422 nigel 75 \fBpcrecallout\fP
2423 nigel 63 .\"
2424     documentation.
2425 nigel 93 .
2426     .
2427 ph10 510 .\" HTML <a name="backtrackcontrol"></a>
2429 ph10 210 .rs
2430     .sp
2431 ph10 211 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2432 ph10 210 are described in the Perl documentation as "experimental and subject to change
2433 ph10 211 or removal in a future version of Perl". It goes on to say: "Their usage in
2434     production code should be noted to avoid problems during upgrades." The same
2435 ph10 210 remarks apply to the PCRE features described in this section.
2436     .P
2437 ph10 341 Since these verbs are specifically related to backtracking, most of them can be
2438     used only when the pattern is to be matched using \fBpcre_exec()\fP, which uses
2439 ph10 345 a backtracking algorithm. With the exception of (*FAIL), which behaves like a
2440 ph10 341 failing negative assertion, they cause an error if encountered by
2441 ph10 210 \fBpcre_dfa_exec()\fP.
2442     .P
2443 ph10 469 If any of these verbs are used in an assertion or subroutine subpattern
2444 ph10 464 (including recursive subpatterns), their effect is confined to that subpattern;
2445     it does not extend to the surrounding pattern. Note that such subpatterns are
2446     processed as anchored at the point where they are tested.
2447 ph10 445 .P
2448 ph10 211 The new verbs make use of what was previously invalid syntax: an opening
2449 ph10 510 parenthesis followed by an asterisk. They are generally of the form
2450 ph10 512 (*VERB) or (*VERB:NAME). Some may take either form, with differing behaviour,
2451 ph10 510 depending on whether or not an argument is present. An name is a sequence of
2452     letters, digits, and underscores. If the name is empty, that is, if the closing
2453     parenthesis immediately follows the colon, the effect is as if the colon were
2454     not there. Any number of these verbs may occur in a pattern.
2455     .P
2456 ph10 512 PCRE contains some optimizations that are used to speed up matching by running
2457     some checks at the start of each match attempt. For example, it may know the
2458     minimum length of matching subject, or that a particular character must be
2459     present. When one of these optimizations suppresses the running of a match, any
2460     included backtracking verbs will not, of course, be processed. You can suppress
2461     the start-of-match optimizations by setting the PCRE_NO_START_OPTIMIZE option
2462 ph10 510 when calling \fBpcre_exec()\fP.
2463 ph10 210 .
2464 ph10 510 .
2465 ph10 210 .SS "Verbs that act immediately"
2466     .rs
2467     .sp
2468 ph10 512 The following verbs act as soon as they are encountered. They may not be
2469 ph10 510 followed by a name.
2470 ph10 210 .sp
2471     (*ACCEPT)
2472     .sp
2473     This verb causes the match to end successfully, skipping the remainder of the
2474 ph10 211 pattern. When inside a recursion, only the innermost pattern is ended
2475 ph10 456 immediately. If (*ACCEPT) is inside capturing parentheses, the data so far is
2476     captured. (This feature was added to PCRE at release 8.00.) For example:
2477 ph10 210 .sp
2478 ph10 447 A((?:A|B(*ACCEPT)|C)D)
2479 ph10 210 .sp
2480 ph10 461 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is captured by
2481 ph10 447 the outer parentheses.
2482 ph10 210 .sp
2483     (*FAIL) or (*F)
2484     .sp
2485 ph10 211 This verb causes the match to fail, forcing backtracking to occur. It is
2486 ph10 210 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2487     probably useful only when combined with (?{}) or (??{}). Those are, of course,
2488     Perl features that are not present in PCRE. The nearest equivalent is the
2489     callout feature, as for example in this pattern:
2490     .sp
2491     a+(?C)(*FAIL)
2492     .sp
2493 ph10 211 A match with the string "aaaa" always fails, but the callout is taken before
2494     each backtrack happens (in this example, 10 times).
2495 ph10 210 .
2496 ph10 510 .
2497     .SS "Recording which path was taken"
2498     .rs
2499     .sp
2500 ph10 512 There is one verb whose main purpose is to track how a match was arrived at,
2501     though it also has a secondary use in conjunction with advancing the match
2502 ph10 510 starting point (see (*SKIP) below).
2503     .sp
2504     (*MARK:NAME) or (*:NAME)
2505     .sp
2506     A name is always required with this verb. There may be as many instances of
2507     (*MARK) as you like in a pattern, and their names do not have to be unique.
2508     .P
2509     When a match succeeds, the name of the last-encountered (*MARK) is passed back
2510     to the caller via the \fIpcre_extra\fP data structure, as described in the
2511     .\" HTML <a href="pcreapi.html#extradata">
2512     .\" </a>
2513     section on \fIpcre_extra\fP
2514     .\"
2515 ph10 512 in the
2516 ph10 510 .\" HREF
2517     \fBpcreapi\fP
2518     .\"
2519     documentation. No data is returned for a partial match. Here is an example of
2520     \fBpcretest\fP output, where the /K modifier requests the retrieval and
2521     outputting of (*MARK) data:
2522     .sp
2523     /X(*MARK:A)Y|X(*MARK:B)Z/K
2524     XY
2525     0: XY
2526     MK: A
2527     XZ
2528     0: XZ
2529     MK: B
2530     .sp
2531 ph10 512 The (*MARK) name is tagged with "MK:" in this output, and in this example it
2532     indicates which of the two alternatives matched. This is a more efficient way
2533 ph10 510 of obtaining this information than putting each alternative in its own
2534     capturing parentheses.
2535     .P
2536     A name may also be returned after a failed match if the final path through the
2537     pattern involves (*MARK). However, unless (*MARK) used in conjunction with
2538     (*COMMIT), this is unlikely to happen for an unanchored pattern because, as the
2539     starting point for matching is advanced, the final check is often with an empty
2540     string, causing a failure before (*MARK) is reached. For example:
2541     .sp
2542     /X(*MARK:A)Y|X(*MARK:B)Z/K
2543     XP
2544     No match
2545     .sp
2546     There are three potential starting points for this match (starting with X,
2547 ph10 512 starting with P, and with an empty string). If the pattern is anchored, the
2548 ph10 510 result is different:
2549     .sp
2550     /^X(*MARK:A)Y|^X(*MARK:B)Z/K
2551     XP
2552     No match, mark = B
2553     .sp
2554 ph10 512 PCRE's start-of-match optimizations can also interfere with this. For example,
2555     if, as a result of a call to \fBpcre_study()\fP, it knows the minimum
2556 ph10 510 subject length for a match, a shorter subject will not be scanned at all.
2557     .P
2558 ph10 512 Note that similar anomalies (though different in detail) exist in Perl, no
2559     doubt for the same reasons. The use of (*MARK) data after a failed match of an
2560 ph10 510 unanchored pattern is not recommended, unless (*COMMIT) is involved.
2561     .
2562     .
2563 ph10 210 .SS "Verbs that act after backtracking"
2564     .rs
2565     .sp
2566 ph10 211 The following verbs do nothing when they are encountered. Matching continues
2567 ph10 510 with what follows, but if there is no subsequent match, causing a backtrack to
2568     the verb, a failure is forced. That is, backtracking cannot pass to the left of
2569     the verb. However, when one of these verbs appears inside an atomic group, its
2570     effect is confined to that group, because once the group has been matched,
2571     there is never any backtracking into it. In this situation, backtracking can
2572 ph10 512 "jump back" to the left of the entire atomic group. (Remember also, as stated
2573 ph10 510 above, that this localization also applies in subroutine calls and assertions.)
2574     .P
2575     These verbs differ in exactly what kind of failure occurs when backtracking
2576     reaches them.
2577 ph10 210 .sp
2578     (*COMMIT)
2579     .sp
2580 ph10 510 This verb, which may not be followed by a name, causes the whole match to fail
2581     outright if the rest of the pattern does not match. Even if the pattern is
2582     unanchored, no further attempts to find a match by advancing the starting point
2583     take place. Once (*COMMIT) has been passed, \fBpcre_exec()\fP is committed to
2584     finding a match at the current starting point, or not at all. For example:
2585 ph10 210 .sp
2586     a+(*COMMIT)b
2587     .sp
2588 ph10 211 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2589 ph10 512 dynamic anchor, or "I've started, so I must finish." The name of the most
2590     recently passed (*MARK) in the path is passed back when (*COMMIT) forces a
2591 ph10 510 match failure.
2592     .P
2593 ph10 512 Note that (*COMMIT) at the start of a pattern is not the same as an anchor,
2594     unless PCRE's start-of-match optimizations are turned off, as shown in this
2595 ph10 510 \fBpcretest\fP example:
2596 ph10 210 .sp
2597 ph10 510 /(*COMMIT)abc/
2598     xyzabc
2599     0: abc
2600     xyzabc\eY
2601     No match
2602 ph10 210 .sp
2603 ph10 512 PCRE knows that any match must start with "a", so the optimization skips along
2604 ph10 510 the subject to "a" before running the first match attempt, which succeeds. When
2605     the optimization is disabled by the \eY escape in the second subject, the match
2606 ph10 512 starts at "x" and so the (*COMMIT) causes it to fail without trying any other
2607 ph10 510 starting points.
2608 ph10 210 .sp
2609 ph10 510 (*PRUNE) or (*PRUNE:NAME)
2610     .sp
2611 ph10 512 This verb causes the match to fail at the current starting position in the
2612 ph10 510 subject if the rest of the pattern does not match. If the pattern is
2613     unanchored, the normal "bumpalong" advance to the next starting character then
2614     happens. Backtracking can occur as usual to the left of (*PRUNE), before it is
2615     reached, or when matching to the right of (*PRUNE), but if there is no match to
2616     the right, backtracking cannot cross (*PRUNE). In simple cases, the use of
2617     (*PRUNE) is just an alternative to an atomic group or possessive quantifier,
2618     but there are some uses of (*PRUNE) that cannot be expressed in any other way.
2619     The behaviour of (*PRUNE:NAME) is the same as (*MARK:NAME)(*PRUNE) when the
2620     match fails completely; the name is passed back if this is the final attempt.
2621     (*PRUNE:NAME) does not pass back a name if the match succeeds. In an anchored
2622     pattern (*PRUNE) has the same effect as (*COMMIT).
2623     .sp
2624 ph10 210 (*SKIP)
2625     .sp
2626 ph10 510 This verb, when given without a name, is like (*PRUNE), except that if the
2627     pattern is unanchored, the "bumpalong" advance is not to the next character,
2628     but to the position in the subject where (*SKIP) was encountered. (*SKIP)
2629     signifies that whatever text was matched leading up to it cannot be part of a
2630     successful match. Consider:
2631 ph10 210 .sp
2632     a+(*SKIP)b
2633     .sp
2634 ph10 211 If the subject is "aaaac...", after the first match attempt fails (starting at
2635 ph10 210 the first character in the string), the starting point skips on to start the
2636 ph10 211 next attempt at "c". Note that a possessive quantifer does not have the same
2637 ph10 456 effect as this example; although it would suppress backtracking during the
2638 ph10 210 first match attempt, the second attempt would start at the second character
2639     instead of skipping on to "c".
2640     .sp
2641 ph10 510 (*SKIP:NAME)
2642 ph10 211 .sp
2643 ph10 512 When (*SKIP) has an associated name, its behaviour is modified. If the
2644     following pattern fails to match, the previous path through the pattern is
2645     searched for the most recent (*MARK) that has the same name. If one is found,
2646     the "bumpalong" advance is to the subject position that corresponds to that
2647     (*MARK) instead of to where (*SKIP) was encountered. If no (*MARK) with a
2648     matching name is found, normal "bumpalong" of one character happens (the
2649 ph10 510 (*SKIP) is ignored).
2650     .sp
2651     (*THEN) or (*THEN:NAME)
2652     .sp
2653 ph10 550 This verb causes a skip to the next alternation in the innermost enclosing
2654     group if the rest of the pattern does not match. That is, it cancels pending
2655     backtracking, but only within the current alternation. Its name comes from the
2656     observation that it can be used for a pattern-based if-then-else block:
2657 ph10 210 .sp
2658     ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
2659     .sp
2660 ph10 211 If the COND1 pattern matches, FOO is tried (and possibly further items after
2661 ph10 210 the end of the group if FOO succeeds); on failure the matcher skips to the
2662 ph10 510 second alternative and tries COND2, without backtracking into COND1. The
2663     behaviour of (*THEN:NAME) is exactly the same as (*MARK:NAME)(*THEN) if the
2664     overall match fails. If (*THEN) is not directly inside an alternation, it acts
2665     like (*PRUNE).
2666 ph10 210 .
2667 ph10 551 .P
2668     The above verbs provide four different "strengths" of control when subsequent
2669     matching fails. (*THEN) is the weakest, carrying on the match at the next
2670     alternation. (*PRUNE) comes next, failing the match at the current starting
2671     position, but allowing an advance to the next character (for an unanchored
2672     pattern). (*SKIP) is similar, except that the advance may be more than one
2673     character. (*COMMIT) is the strongest, causing the entire match to fail.
2674     .P
2675     If more than one is present in a pattern, the "stongest" one wins. For example,
2676     consider this pattern, where A, B, etc. are complex pattern fragments:
2677     .sp
2678     (A(*COMMIT)B(*THEN)C|D)
2679     .sp
2680     Once A has matched, PCRE is committed to this match, at the current starting
2681     position. If subsequently B matches, but C does not, the normal (*THEN) action
2682     of trying the next alternation (that is, D) does not happen because (*COMMIT)
2683     overrides.
2684 ph10 210 .
2685 ph10 551 .
2686 nigel 93 .SH "SEE ALSO"
2687     .rs
2688     .sp
2689 ph10 461 \fBpcreapi\fP(3), \fBpcrecallout\fP(3), \fBpcrematching\fP(3),
2690 ph10 456 \fBpcresyntax\fP(3), \fBpcre\fP(3).
2691 ph10 99 .
2692     .
2693     .SH AUTHOR
2694     .rs
2695     .sp
2696     .nf
2697     Philip Hazel
2698     University Computing Service
2699     Cambridge CB2 3QH, England.
2700     .fi
2701     .
2702     .
2703     .SH REVISION
2704     .rs
2705     .sp
2706     .nf
2707 ph10 555 Last updated: 26 October 2010
2708 ph10 488 Copyright (c) 1997-2010 University of Cambridge.
2709 ph10 99 .fi


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

ViewVC Help
Powered by ViewVC 1.1.12