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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 518 - (hide annotations) (download)
Tue May 18 15:47:01 2010 UTC (4 years, 2 months ago) by ph10
File size: 109479 byte(s)
Added PCRE_UCP and related stuff to make \w etc use Unicode properties.

1 nigel 79 .TH PCREPATTERN 3
2 nigel 63 .SH NAME
3     PCRE - Perl-compatible regular expressions
4 nigel 75 .SH "PCRE REGULAR EXPRESSION DETAILS"
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 518 Another special sequence that may appear at the start of a pattern or in
46     combination with (*UTF8) is:
47     .sp
48     (*UCP)
49     .sp
50     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     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 227 .SH "NEWLINE CONVENTIONS"
70     .rs
71     .sp
72     PCRE supports five different conventions for indicating line breaks in
73     strings: a single CR (carriage return) character, a single LF (linefeed)
74     character, the two-character sequence CRLF, any of the three preceding, or any
75     Unicode newline sequence. The
76     .\" HREF
77     \fBpcreapi\fP
78     .\"
79     page has
80     .\" HTML <a href="pcreapi.html#newlines">
81     .\" </a>
82     further discussion
83     .\"
84     about newlines, and shows how to set the newline convention in the
85     \fIoptions\fP arguments for the compiling and matching functions.
86     .P
87     It is also possible to specify a newline convention by starting a pattern
88     string with one of the following five sequences:
89     .sp
90     (*CR) carriage return
91     (*LF) linefeed
92     (*CRLF) carriage return, followed by linefeed
93     (*ANYCRLF) any of the three above
94     (*ANY) all Unicode newline sequences
95     .sp
96 ph10 461 These override the default and the options given to \fBpcre_compile()\fP or
97 ph10 456 \fBpcre_compile2()\fP. For example, on a Unix system where LF is the default
98     newline sequence, the pattern
99 ph10 227 .sp
100     (*CR)a.b
101     .sp
102     changes the convention to CR. That pattern matches "a\enb" because LF is no
103     longer a newline. Note that these special settings, which are not
104     Perl-compatible, are recognized only at the very start of a pattern, and that
105 ph10 231 they must be in upper case. If more than one of them is present, the last one
106     is used.
107     .P
108 ph10 514 The newline convention affects the interpretation of the dot metacharacter when
109     PCRE_DOTALL is not set, and also the behaviour of \eN. However, it does not
110     affect what the \eR escape sequence matches. By default, this is any Unicode
111     newline sequence, for Perl compatibility. However, this can be changed; see the
112     description of \eR in the section entitled
113 ph10 231 .\" HTML <a href="#newlineseq">
114     .\" </a>
115     "Newline sequences"
116     .\"
117 ph10 247 below. A change of \eR setting can be combined with a change of newline
118 ph10 246 convention.
119 ph10 227 .
120     .
121 nigel 93 .SH "CHARACTERS AND METACHARACTERS"
122     .rs
123     .sp
124 nigel 63 A regular expression is a pattern that is matched against a subject string from
125     left to right. Most characters stand for themselves in a pattern, and match the
126     corresponding characters in the subject. As a trivial example, the pattern
127 nigel 75 .sp
128 nigel 63 The quick brown fox
129 nigel 75 .sp
130 nigel 77 matches a portion of a subject string that is identical to itself. When
131     caseless matching is specified (the PCRE_CASELESS option), letters are matched
132     independently of case. In UTF-8 mode, PCRE always understands the concept of
133     case for characters whose values are less than 128, so caseless matching is
134     always possible. For characters with higher values, the concept of case is
135     supported if PCRE is compiled with Unicode property support, but not otherwise.
136     If you want to use caseless matching for characters 128 and above, you must
137     ensure that PCRE is compiled with Unicode property support as well as with
138     UTF-8 support.
139     .P
140     The power of regular expressions comes from the ability to include alternatives
141     and repetitions in the pattern. These are encoded in the pattern by the use of
142 nigel 75 \fImetacharacters\fP, which do not stand for themselves but instead are
143 nigel 63 interpreted in some special way.
144 nigel 75 .P
145     There are two different sets of metacharacters: those that are recognized
146 nigel 63 anywhere in the pattern except within square brackets, and those that are
147 nigel 93 recognized within square brackets. Outside square brackets, the metacharacters
148     are as follows:
149 nigel 75 .sp
150     \e general escape character with several uses
151 nigel 63 ^ assert start of string (or line, in multiline mode)
152     $ assert end of string (or line, in multiline mode)
153     . match any character except newline (by default)
154     [ start character class definition
155     | start of alternative branch
156     ( start subpattern
157     ) end subpattern
158     ? extends the meaning of (
159     also 0 or 1 quantifier
160     also quantifier minimizer
161     * 0 or more quantifier
162     + 1 or more quantifier
163     also "possessive quantifier"
164     { start min/max quantifier
165 nigel 75 .sp
166 nigel 63 Part of a pattern that is in square brackets is called a "character class". In
167 nigel 75 a character class the only metacharacters are:
168     .sp
169     \e general escape character
170 nigel 63 ^ negate the class, but only if the first character
171     - indicates character range
172 nigel 75 .\" JOIN
173 nigel 63 [ POSIX character class (only if followed by POSIX
174     syntax)
175     ] terminates the character class
176 nigel 75 .sp
177     The following sections describe the use of each of the metacharacters.
178     .
179 nigel 93 .
180 nigel 63 .SH BACKSLASH
181     .rs
182     .sp
183     The backslash character has several uses. Firstly, if it is followed by a
184 nigel 91 non-alphanumeric character, it takes away any special meaning that character
185     may have. This use of backslash as an escape character applies both inside and
186 nigel 63 outside character classes.
187 nigel 75 .P
188     For example, if you want to match a * character, you write \e* in the pattern.
189 nigel 63 This escaping action applies whether or not the following character would
190 nigel 75 otherwise be interpreted as a metacharacter, so it is always safe to precede a
191     non-alphanumeric with backslash to specify that it stands for itself. In
192     particular, if you want to match a backslash, you write \e\e.
193     .P
194 nigel 63 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
195     pattern (other than in a character class) and characters between a # outside
196 nigel 91 a character class and the next newline are ignored. An escaping backslash can
197     be used to include a whitespace or # character as part of the pattern.
198 nigel 75 .P
199 nigel 63 If you want to remove the special meaning from a sequence of characters, you
200 nigel 75 can do so by putting them between \eQ and \eE. This is different from Perl in
201     that $ and @ are handled as literals in \eQ...\eE sequences in PCRE, whereas in
202 nigel 63 Perl, $ and @ cause variable interpolation. Note the following examples:
203 nigel 75 .sp
204 nigel 63 Pattern PCRE matches Perl matches
205 nigel 75 .sp
206     .\" JOIN
207     \eQabc$xyz\eE abc$xyz abc followed by the
208 nigel 63 contents of $xyz
209 nigel 75 \eQabc\e$xyz\eE abc\e$xyz abc\e$xyz
210     \eQabc\eE\e$\eQxyz\eE abc$xyz abc$xyz
211     .sp
212     The \eQ...\eE sequence is recognized both inside and outside character classes.
213     .
214     .
215     .\" HTML <a name="digitsafterbackslash"></a>
216     .SS "Non-printing characters"
217     .rs
218     .sp
219 nigel 63 A second use of backslash provides a way of encoding non-printing characters
220     in patterns in a visible manner. There is no restriction on the appearance of
221     non-printing characters, apart from the binary zero that terminates a pattern,
222 ph10 456 but when a pattern is being prepared by text editing, it is often easier to use
223     one of the following escape sequences than the binary character it represents:
224 nigel 75 .sp
225     \ea alarm, that is, the BEL character (hex 07)
226     \ecx "control-x", where x is any character
227     \ee escape (hex 1B)
228     \ef formfeed (hex 0C)
229 ph10 227 \en linefeed (hex 0A)
230 nigel 75 \er carriage return (hex 0D)
231     \et tab (hex 09)
232 ph10 488 \eddd character with octal code ddd, or back reference
233 nigel 75 \exhh character with hex code hh
234 nigel 87 \ex{hhh..} character with hex code hhh..
235 nigel 75 .sp
236     The precise effect of \ecx is as follows: if x is a lower case letter, it
237 nigel 63 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
238 nigel 75 Thus \ecz becomes hex 1A, but \ec{ becomes hex 3B, while \ec; becomes hex
239 nigel 63 7B.
240 nigel 75 .P
241     After \ex, from zero to two hexadecimal digits are read (letters can be in
242 nigel 87 upper or lower case). Any number of hexadecimal digits may appear between \ex{
243     and }, but the value of the character code must be less than 256 in non-UTF-8
244 ph10 211 mode, and less than 2**31 in UTF-8 mode. That is, the maximum value in
245     hexadecimal is 7FFFFFFF. Note that this is bigger than the largest Unicode code
246     point, which is 10FFFF.
247 nigel 75 .P
248 ph10 211 If characters other than hexadecimal digits appear between \ex{ and }, or if
249     there is no terminating }, this form of escape is not recognized. Instead, the
250     initial \ex will be interpreted as a basic hexadecimal escape, with no
251     following digits, giving a character whose value is zero.
252     .P
253 nigel 63 Characters whose value is less than 256 can be defined by either of the two
254 nigel 87 syntaxes for \ex. There is no difference in the way they are handled. For
255     example, \exdc is exactly the same as \ex{dc}.
256 nigel 75 .P
257 nigel 91 After \e0 up to two further octal digits are read. If there are fewer than two
258     digits, just those that are present are used. Thus the sequence \e0\ex\e07
259     specifies two binary zeros followed by a BEL character (code value 7). Make
260     sure you supply two digits after the initial zero if the pattern character that
261     follows is itself an octal digit.
262 nigel 75 .P
263 nigel 63 The handling of a backslash followed by a digit other than 0 is complicated.
264     Outside a character class, PCRE reads it and any following digits as a decimal
265     number. If the number is less than 10, or if there have been at least that many
266     previous capturing left parentheses in the expression, the entire sequence is
267 nigel 75 taken as a \fIback reference\fP. A description of how this works is given
268     .\" HTML <a href="#backreferences">
269     .\" </a>
270     later,
271     .\"
272     following the discussion of
273     .\" HTML <a href="#subpattern">
274     .\" </a>
275     parenthesized subpatterns.
276     .\"
277     .P
278 nigel 63 Inside a character class, or if the decimal number is greater than 9 and there
279     have not been that many capturing subpatterns, PCRE re-reads up to three octal
280 nigel 93 digits following the backslash, and uses them to generate a data character. Any
281 nigel 91 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
282     character specified in octal must be less than \e400. In UTF-8 mode, values up
283     to \e777 are permitted. For example:
284 nigel 75 .sp
285     \e040 is another way of writing a space
286     .\" JOIN
287     \e40 is the same, provided there are fewer than 40
288 nigel 63 previous capturing subpatterns
289 nigel 75 \e7 is always a back reference
290     .\" JOIN
291     \e11 might be a back reference, or another way of
292 nigel 63 writing a tab
293 nigel 75 \e011 is always a tab
294     \e0113 is a tab followed by the character "3"
295     .\" JOIN
296     \e113 might be a back reference, otherwise the
297 nigel 63 character with octal code 113
298 nigel 75 .\" JOIN
299     \e377 might be a back reference, otherwise
300 nigel 63 the byte consisting entirely of 1 bits
301 nigel 75 .\" JOIN
302     \e81 is either a back reference, or a binary zero
303 nigel 63 followed by the two characters "8" and "1"
304 nigel 75 .sp
305 nigel 63 Note that octal values of 100 or greater must not be introduced by a leading
306     zero, because no more than three octal digits are ever read.
307 nigel 75 .P
308 nigel 91 All the sequences that define a single character value can be used both inside
309     and outside character classes. In addition, inside a character class, the
310 ph10 513 sequence \eb is interpreted as the backspace character (hex 08). The sequences
311 ph10 514 \eB, \eN, \eR, and \eX are not special inside a character class. Like any other
312 ph10 513 unrecognized escape sequences, they are treated as the literal characters "B",
313 ph10 514 "N", "R", and "X" by default, but cause an error if the PCRE_EXTRA option is
314     set. Outside a character class, these sequences have different meanings.
315 nigel 75 .
316     .
317 nigel 93 .SS "Absolute and relative back references"
318     .rs
319     .sp
320 ph10 208 The sequence \eg followed by an unsigned or a negative number, optionally
321     enclosed in braces, is an absolute or relative back reference. A named back
322     reference can be coded as \eg{name}. Back references are discussed
323 nigel 93 .\" HTML <a href="#backreferences">
324     .\" </a>
325     later,
326     .\"
327     following the discussion of
328     .\" HTML <a href="#subpattern">
329     .\" </a>
330     parenthesized subpatterns.
331     .\"
332     .
333     .
334 ph10 333 .SS "Absolute and relative subroutine calls"
335     .rs
336     .sp
337 ph10 345 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
338     a number enclosed either in angle brackets or single quotes, is an alternative
339     syntax for referencing a subpattern as a "subroutine". Details are discussed
340 ph10 333 .\" HTML <a href="#onigurumasubroutines">
341     .\" </a>
342     later.
343     .\"
344 ph10 345 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
345 ph10 461 synonymous. The former is a back reference; the latter is a
346 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
347     .\" </a>
348     subroutine
349     .\"
350     call.
351 ph10 333 .
352     .
353 ph10 518 .\" HTML <a name="genericchartypes"></a>
354 nigel 75 .SS "Generic character types"
355     .rs
356     .sp
357 ph10 514 Another use of backslash is for specifying generic character types:
358 nigel 75 .sp
359 ph10 182 \ed any decimal digit
360 nigel 75 \eD any character that is not a decimal digit
361 ph10 178 \eh any horizontal whitespace character
362 ph10 182 \eH any character that is not a horizontal whitespace character
363 nigel 75 \es any whitespace character
364     \eS any character that is not a whitespace character
365 ph10 178 \ev any vertical whitespace character
366 ph10 182 \eV any character that is not a vertical whitespace character
367 nigel 75 \ew any "word" character
368     \eW any "non-word" character
369     .sp
370 ph10 514 There is also the single sequence \eN, which matches a non-newline character.
371     This is the same as
372     .\" HTML <a href="#fullstopdot">
373     .\" </a>
374     the "." metacharacter
375     .\"
376     when PCRE_DOTALL is not set.
377 nigel 75 .P
378 ph10 514 Each pair of lower and upper case escape sequences partitions the complete set
379     of characters into two disjoint sets. Any given character matches one, and only
380 ph10 518 one, of each pair. The sequences can appear both inside and outside character
381 nigel 75 classes. They each match one character of the appropriate type. If the current
382 ph10 518 matching point is at the end of the subject string, all of them fail, because
383 nigel 75 there is no character to match.
384     .P
385     For compatibility with Perl, \es does not match the VT character (code 11).
386     This makes it different from the the POSIX "space" class. The \es characters
387 ph10 178 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
388 nigel 91 included in a Perl script, \es may match the VT character. In PCRE, it never
389 ph10 178 does.
390 nigel 75 .P
391 ph10 518 A "word" character is an underscore or any character that is a letter or digit.
392     By default, the definition of letters and digits is controlled by PCRE's
393     low-valued character tables, and may vary if locale-specific matching is taking
394     place (see
395     .\" HTML <a href="pcreapi.html#localesupport">
396     .\" </a>
397     "Locale support"
398     .\"
399     in the
400     .\" HREF
401     \fBpcreapi\fP
402     .\"
403     page). For example, in a French locale such as "fr_FR" in Unix-like systems,
404     or "french" in Windows, some character codes greater than 128 are used for
405     accented letters, and these are then matched by \ew. The use of locales with
406     Unicode is discouraged.
407 ph10 178 .P
408 ph10 518 By default, in UTF-8 mode, characters with values greater than 128 never match
409     \ed, \es, or \ew, and always match \eD, \eS, and \eW. These sequences retain
410     their original meanings from before UTF-8 support was available, mainly for
411     efficiency reasons. However, if PCRE is compiled with Unicode property support,
412     and the PCRE_UCP option is set, the behaviour is changed so that Unicode
413     properties are used to determine character types, as follows:
414     .sp
415     \ed any character that \ep{Nd} matches (decimal digit)
416     \es any character that \ep{Z} matches, plus HT, LF, FF, CR
417     \ew any character that \ep{L} or \ep{N} matches, plus underscore
418     .sp
419     The upper case escapes match the inverse sets of characters. Note that \ed
420     matches only decimal digits, whereas \ew matches any Unicode digit, as well as
421     any Unicode letter, and underscore. Note also that PCRE_UCP affects \eb, and
422     \eB because they are defined in terms of \ew and \eW. Matching these sequences
423     is noticeably slower when PCRE_UCP is set.
424     .P
425 ph10 182 The sequences \eh, \eH, \ev, and \eV are Perl 5.10 features. In contrast to the
426 ph10 518 other sequences, which match only ASCII characters by default, these always
427     match certain high-valued codepoints in UTF-8 mode, whether or not PCRE_UCP is
428     set. The horizontal space characters are:
429 ph10 178 .sp
430     U+0009 Horizontal tab
431     U+0020 Space
432     U+00A0 Non-break space
433     U+1680 Ogham space mark
434     U+180E Mongolian vowel separator
435     U+2000 En quad
436     U+2001 Em quad
437     U+2002 En space
438     U+2003 Em space
439     U+2004 Three-per-em space
440     U+2005 Four-per-em space
441     U+2006 Six-per-em space
442     U+2007 Figure space
443     U+2008 Punctuation space
444     U+2009 Thin space
445     U+200A Hair space
446     U+202F Narrow no-break space
447     U+205F Medium mathematical space
448     U+3000 Ideographic space
449     .sp
450     The vertical space characters are:
451     .sp
452     U+000A Linefeed
453     U+000B Vertical tab
454     U+000C Formfeed
455     U+000D Carriage return
456     U+0085 Next line
457     U+2028 Line separator
458     U+2029 Paragraph separator
459 nigel 75 .
460     .
461 ph10 231 .\" HTML <a name="newlineseq"></a>
462 nigel 93 .SS "Newline sequences"
463     .rs
464     .sp
465 ph10 231 Outside a character class, by default, the escape sequence \eR matches any
466     Unicode newline sequence. This is a Perl 5.10 feature. In non-UTF-8 mode \eR is
467     equivalent to the following:
468 nigel 93 .sp
469     (?>\er\en|\en|\ex0b|\ef|\er|\ex85)
470     .sp
471     This is an example of an "atomic group", details of which are given
472     .\" HTML <a href="#atomicgroup">
473     .\" </a>
474     below.
475     .\"
476     This particular group matches either the two-character sequence CR followed by
477     LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
478     U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
479     line, U+0085). The two-character sequence is treated as a single unit that
480     cannot be split.
481     .P
482     In UTF-8 mode, two additional characters whose codepoints are greater than 255
483     are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
484     Unicode character property support is not needed for these characters to be
485     recognized.
486     .P
487 ph10 231 It is possible to restrict \eR to match only CR, LF, or CRLF (instead of the
488     complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
489 ph10 247 either at compile time or when the pattern is matched. (BSR is an abbrevation
490 ph10 246 for "backslash R".) This can be made the default when PCRE is built; if this is
491     the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
492     It is also possible to specify these settings by starting a pattern string with
493     one of the following sequences:
494 ph10 231 .sp
495     (*BSR_ANYCRLF) CR, LF, or CRLF only
496     (*BSR_UNICODE) any Unicode newline sequence
497     .sp
498 ph10 461 These override the default and the options given to \fBpcre_compile()\fP or
499 ph10 456 \fBpcre_compile2()\fP, but they can be overridden by options given to
500     \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP. Note that these special settings,
501     which are not Perl-compatible, are recognized only at the very start of a
502     pattern, and that they must be in upper case. If more than one of them is
503     present, the last one is used. They can be combined with a change of newline
504 ph10 518 convention; for example, a pattern can start with:
505 ph10 246 .sp
506     (*ANY)(*BSR_ANYCRLF)
507     .sp
508 ph10 518 They can also be combined with the (*UTF8) or (*UCP) special sequences. Inside
509     a character class, \eR is treated as an unrecognized escape sequence, and so
510     matches the letter "R" by default, but causes an error if PCRE_EXTRA is set.
511 nigel 93 .
512     .
513 nigel 75 .\" HTML <a name="uniextseq"></a>
514     .SS Unicode character properties
515     .rs
516     .sp
517     When PCRE is built with Unicode character property support, three additional
518 ph10 184 escape sequences that match characters with specific properties are available.
519     When not in UTF-8 mode, these sequences are of course limited to testing
520     characters whose codepoints are less than 256, but they do work in this mode.
521     The extra escape sequences are:
522 nigel 75 .sp
523 nigel 87 \ep{\fIxx\fP} a character with the \fIxx\fP property
524     \eP{\fIxx\fP} a character without the \fIxx\fP property
525     \eX an extended Unicode sequence
526 nigel 75 .sp
527 nigel 87 The property names represented by \fIxx\fP above are limited to the Unicode
528 ph10 517 script names, the general category properties, "Any", which matches any
529     character (including newline), and some special PCRE properties (described
530     in the
531     .\" HTML <a href="#extraprops">
532     .\" </a>
533     next section).
534     .\"
535     Other Perl properties such as "InMusicalSymbols" are not currently supported by
536     PCRE. Note that \eP{Any} does not match any characters, so always causes a
537     match failure.
538 nigel 75 .P
539 nigel 87 Sets of Unicode characters are defined as belonging to certain scripts. A
540     character from one of these sets can be matched using a script name. For
541     example:
542 nigel 75 .sp
543 nigel 87 \ep{Greek}
544     \eP{Han}
545     .sp
546     Those that are not part of an identified script are lumped together as
547     "Common". The current list of scripts is:
548     .P
549     Arabic,
550     Armenian,
551 ph10 491 Avestan,
552 nigel 93 Balinese,
553 ph10 491 Bamum,
554 nigel 87 Bengali,
555     Bopomofo,
556     Braille,
557     Buginese,
558     Buhid,
559     Canadian_Aboriginal,
560 ph10 491 Carian,
561     Cham,
562 nigel 87 Cherokee,
563     Common,
564     Coptic,
565 nigel 93 Cuneiform,
566 nigel 87 Cypriot,
567     Cyrillic,
568     Deseret,
569     Devanagari,
570 ph10 491 Egyptian_Hieroglyphs,
571 nigel 87 Ethiopic,
572     Georgian,
573     Glagolitic,
574     Gothic,
575     Greek,
576     Gujarati,
577     Gurmukhi,
578     Han,
579     Hangul,
580     Hanunoo,
581     Hebrew,
582     Hiragana,
583 ph10 491 Imperial_Aramaic,
584 nigel 87 Inherited,
585 ph10 491 Inscriptional_Pahlavi,
586     Inscriptional_Parthian,
587     Javanese,
588     Kaithi,
589 nigel 87 Kannada,
590     Katakana,
591 ph10 491 Kayah_Li,
592 nigel 87 Kharoshthi,
593     Khmer,
594     Lao,
595     Latin,
596 ph10 491 Lepcha,
597 nigel 87 Limbu,
598     Linear_B,
599 ph10 491 Lisu,
600     Lycian,
601     Lydian,
602 nigel 87 Malayalam,
603 ph10 491 Meetei_Mayek,
604 nigel 87 Mongolian,
605     Myanmar,
606     New_Tai_Lue,
607 nigel 93 Nko,
608 nigel 87 Ogham,
609     Old_Italic,
610     Old_Persian,
611 ph10 491 Old_South_Arabian,
612     Old_Turkic,
613     Ol_Chiki,
614 nigel 87 Oriya,
615     Osmanya,
616 nigel 93 Phags_Pa,
617     Phoenician,
618 ph10 491 Rejang,
619 nigel 87 Runic,
620 ph10 491 Samaritan,
621     Saurashtra,
622 nigel 87 Shavian,
623     Sinhala,
624 ph10 491 Sundanese,
625 nigel 87 Syloti_Nagri,
626     Syriac,
627     Tagalog,
628     Tagbanwa,
629     Tai_Le,
630 ph10 491 Tai_Tham,
631     Tai_Viet,
632 nigel 87 Tamil,
633     Telugu,
634     Thaana,
635     Thai,
636     Tibetan,
637     Tifinagh,
638     Ugaritic,
639 ph10 491 Vai,
640 nigel 87 Yi.
641     .P
642 ph10 517 Each character has exactly one Unicode general category property, specified by
643     a two-letter abbreviation. For compatibility with Perl, negation can be
644     specified by including a circumflex between the opening brace and the property
645     name. For example, \ep{^Lu} is the same as \eP{Lu}.
646 nigel 87 .P
647     If only one letter is specified with \ep or \eP, it includes all the general
648     category properties that start with that letter. In this case, in the absence
649     of negation, the curly brackets in the escape sequence are optional; these two
650     examples have the same effect:
651     .sp
652 nigel 75 \ep{L}
653     \epL
654     .sp
655 nigel 87 The following general category property codes are supported:
656 nigel 75 .sp
657     C Other
658     Cc Control
659     Cf Format
660     Cn Unassigned
661     Co Private use
662     Cs Surrogate
663     .sp
664     L Letter
665     Ll Lower case letter
666     Lm Modifier letter
667     Lo Other letter
668     Lt Title case letter
669     Lu Upper case letter
670     .sp
671     M Mark
672     Mc Spacing mark
673     Me Enclosing mark
674     Mn Non-spacing mark
675     .sp
676     N Number
677     Nd Decimal number
678     Nl Letter number
679     No Other number
680     .sp
681     P Punctuation
682     Pc Connector punctuation
683     Pd Dash punctuation
684     Pe Close punctuation
685     Pf Final punctuation
686     Pi Initial punctuation
687     Po Other punctuation
688     Ps Open punctuation
689     .sp
690     S Symbol
691     Sc Currency symbol
692     Sk Modifier symbol
693     Sm Mathematical symbol
694     So Other symbol
695     .sp
696     Z Separator
697     Zl Line separator
698     Zp Paragraph separator
699     Zs Space separator
700     .sp
701 nigel 87 The special property L& is also supported: it matches a character that has
702     the Lu, Ll, or Lt property, in other words, a letter that is not classified as
703     a modifier or "other".
704 nigel 75 .P
705 ph10 211 The Cs (Surrogate) property applies only to characters in the range U+D800 to
706     U+DFFF. Such characters are not valid in UTF-8 strings (see RFC 3629) and so
707     cannot be tested by PCRE, unless UTF-8 validity checking has been turned off
708     (see the discussion of PCRE_NO_UTF8_CHECK in the
709     .\" HREF
710     \fBpcreapi\fP
711     .\"
712 ph10 451 page). Perl does not support the Cs property.
713 ph10 211 .P
714 ph10 451 The long synonyms for property names that Perl supports (such as \ep{Letter})
715 nigel 91 are not supported by PCRE, nor is it permitted to prefix any of these
716 nigel 87 properties with "Is".
717     .P
718     No character that is in the Unicode table has the Cn (unassigned) property.
719     Instead, this property is assumed for any code point that is not in the
720     Unicode table.
721     .P
722 nigel 75 Specifying caseless matching does not affect these escape sequences. For
723     example, \ep{Lu} always matches only upper case letters.
724     .P
725     The \eX escape matches any number of Unicode characters that form an extended
726     Unicode sequence. \eX is equivalent to
727     .sp
728     (?>\ePM\epM*)
729     .sp
730     That is, it matches a character without the "mark" property, followed by zero
731     or more characters with the "mark" property, and treats the sequence as an
732     atomic group
733     .\" HTML <a href="#atomicgroup">
734     .\" </a>
735     (see below).
736     .\"
737     Characters with the "mark" property are typically accents that affect the
738 ph10 185 preceding character. None of them have codepoints less than 256, so in
739 ph10 184 non-UTF-8 mode \eX matches any one character.
740 nigel 75 .P
741     Matching characters by Unicode property is not fast, because PCRE has to search
742     a structure that contains data for over fifteen thousand characters. That is
743     why the traditional escape sequences such as \ed and \ew do not use Unicode
744 ph10 518 properties in PCRE by default, though you can make them do so by setting the
745     PCRE_UCP option for \fBpcre_compile()\fP or by starting the pattern with
746     (*UCP).
747 nigel 75 .
748     .
749 ph10 517 .\" HTML <a name="extraprops"></a>
750     .SS PCRE's additional properties
751     .rs
752     .sp
753     As well as the standard Unicode properties described in the previous
754     section, PCRE supports four more that make it possible to convert traditional
755     escape sequences such as \ew and \es and POSIX character classes to use Unicode
756 ph10 518 properties. PCRE uses these non-standard, non-Perl properties internally when
757     PCRE_UCP is set. They are:
758 ph10 517 .sp
759     Xan Any alphanumeric character
760     Xps Any POSIX space character
761     Xsp Any Perl space character
762     Xwd Any Perl "word" character
763     .sp
764     Xan matches characters that have either the L (letter) or the N (number)
765     property. Xps matches the characters tab, linefeed, vertical tab, formfeed, or
766     carriage return, and any other character that has the Z (separator) property.
767     Xsp is the same as Xps, except that vertical tab is excluded. Xwd matches the
768     same characters as Xan, plus underscore.
769     .
770     .
771 ph10 168 .\" HTML <a name="resetmatchstart"></a>
772     .SS "Resetting the match start"
773     .rs
774     .sp
775 ph10 172 The escape sequence \eK, which is a Perl 5.10 feature, causes any previously
776     matched characters not to be included in the final matched sequence. For
777 ph10 168 example, the pattern:
778     .sp
779     foo\eKbar
780     .sp
781 ph10 172 matches "foobar", but reports that it has matched "bar". This feature is
782 ph10 168 similar to a lookbehind assertion
783     .\" HTML <a href="#lookbehind">
784     .\" </a>
785     (described below).
786     .\"
787 ph10 172 However, in this case, the part of the subject before the real match does not
788     have to be of fixed length, as lookbehind assertions do. The use of \eK does
789 ph10 168 not interfere with the setting of
790     .\" HTML <a href="#subpattern">
791     .\" </a>
792     captured substrings.
793 ph10 172 .\"
794 ph10 168 For example, when the pattern
795     .sp
796     (foo)\eKbar
797     .sp
798 ph10 172 matches "foobar", the first substring is still set to "foo".
799 ph10 500 .P
800 ph10 507 Perl documents that the use of \eK within assertions is "not well defined". In
801     PCRE, \eK is acted upon when it occurs inside positive assertions, but is
802 ph10 500 ignored in negative assertions.
803 ph10 168 .
804     .
805 nigel 75 .\" HTML <a name="smallassertions"></a>
806     .SS "Simple assertions"
807     .rs
808     .sp
809 nigel 93 The final use of backslash is for certain simple assertions. An assertion
810 nigel 63 specifies a condition that has to be met at a particular point in a match,
811     without consuming any characters from the subject string. The use of
812 nigel 75 subpatterns for more complicated assertions is described
813     .\" HTML <a href="#bigassertions">
814     .\" </a>
815     below.
816     .\"
817 nigel 91 The backslashed assertions are:
818 nigel 75 .sp
819     \eb matches at a word boundary
820     \eB matches when not at a word boundary
821 nigel 93 \eA matches at the start of the subject
822     \eZ matches at the end of the subject
823     also matches before a newline at the end of the subject
824     \ez matches only at the end of the subject
825     \eG matches at the first matching position in the subject
826 nigel 75 .sp
827 ph10 513 Inside a character class, \eb has a different meaning; it matches the backspace
828     character. If any other of these assertions appears in a character class, by
829     default it matches the corresponding literal character (for example, \eB
830     matches the letter B). However, if the PCRE_EXTRA option is set, an "invalid
831     escape sequence" error is generated instead.
832 nigel 75 .P
833 nigel 63 A word boundary is a position in the subject string where the current character
834 nigel 75 and the previous character do not both match \ew or \eW (i.e. one matches
835     \ew and the other matches \eW), or the start or end of the string if the
836 ph10 518 first or last character matches \ew, respectively. In UTF-8 mode, the meanings
837     of \ew and \eW can be changed by setting the PCRE_UCP option. When this is
838     done, it also affects \eb and \eB. Neither PCRE nor Perl has a separate "start
839     of word" or "end of word" metasequence. However, whatever follows \eb normally
840     determines which it is. For example, the fragment \eba matches "a" at the start
841     of a word.
842 nigel 75 .P
843     The \eA, \eZ, and \ez assertions differ from the traditional circumflex and
844     dollar (described in the next section) in that they only ever match at the very
845     start and end of the subject string, whatever options are set. Thus, they are
846     independent of multiline mode. These three assertions are not affected by the
847     PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
848     circumflex and dollar metacharacters. However, if the \fIstartoffset\fP
849     argument of \fBpcre_exec()\fP is non-zero, indicating that matching is to start
850     at a point other than the beginning of the subject, \eA can never match. The
851 nigel 91 difference between \eZ and \ez is that \eZ matches before a newline at the end
852     of the string as well as at the very end, whereas \ez matches only at the end.
853 nigel 75 .P
854     The \eG assertion is true only when the current matching position is at the
855     start point of the match, as specified by the \fIstartoffset\fP argument of
856     \fBpcre_exec()\fP. It differs from \eA when the value of \fIstartoffset\fP is
857     non-zero. By calling \fBpcre_exec()\fP multiple times with appropriate
858 nigel 63 arguments, you can mimic Perl's /g option, and it is in this kind of
859 nigel 75 implementation where \eG can be useful.
860     .P
861     Note, however, that PCRE's interpretation of \eG, as the start of the current
862 nigel 63 match, is subtly different from Perl's, which defines it as the end of the
863     previous match. In Perl, these can be different when the previously matched
864     string was empty. Because PCRE does just one match at a time, it cannot
865     reproduce this behaviour.
866 nigel 75 .P
867     If all the alternatives of a pattern begin with \eG, the expression is anchored
868 nigel 63 to the starting match position, and the "anchored" flag is set in the compiled
869     regular expression.
870 nigel 75 .
871     .
872     .SH "CIRCUMFLEX AND DOLLAR"
873 nigel 63 .rs
874     .sp
875     Outside a character class, in the default matching mode, the circumflex
876 nigel 75 character is an assertion that is true only if the current matching point is
877     at the start of the subject string. If the \fIstartoffset\fP argument of
878     \fBpcre_exec()\fP is non-zero, circumflex can never match if the PCRE_MULTILINE
879 nigel 63 option is unset. Inside a character class, circumflex has an entirely different
880 nigel 75 meaning
881     .\" HTML <a href="#characterclass">
882     .\" </a>
883     (see below).
884     .\"
885     .P
886 nigel 63 Circumflex need not be the first character of the pattern if a number of
887     alternatives are involved, but it should be the first thing in each alternative
888     in which it appears if the pattern is ever to match that branch. If all
889     possible alternatives start with a circumflex, that is, if the pattern is
890     constrained to match only at the start of the subject, it is said to be an
891     "anchored" pattern. (There are also other constructs that can cause a pattern
892     to be anchored.)
893 nigel 75 .P
894     A dollar character is an assertion that is true only if the current matching
895 nigel 63 point is at the end of the subject string, or immediately before a newline
896 nigel 91 at the end of the string (by default). Dollar need not be the last character of
897     the pattern if a number of alternatives are involved, but it should be the last
898     item in any branch in which it appears. Dollar has no special meaning in a
899     character class.
900 nigel 75 .P
901 nigel 63 The meaning of dollar can be changed so that it matches only at the very end of
902     the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
903 nigel 75 does not affect the \eZ assertion.
904     .P
905 nigel 63 The meanings of the circumflex and dollar characters are changed if the
906 nigel 91 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
907     immediately after internal newlines as well as at the start of the subject
908     string. It does not match after a newline that ends the string. A dollar
909     matches before any newlines in the string, as well as at the very end, when
910     PCRE_MULTILINE is set. When newline is specified as the two-character
911     sequence CRLF, isolated CR and LF characters do not indicate newlines.
912 nigel 75 .P
913 nigel 91 For example, the pattern /^abc$/ matches the subject string "def\enabc" (where
914     \en represents a newline) in multiline mode, but not otherwise. Consequently,
915     patterns that are anchored in single line mode because all branches start with
916     ^ are not anchored in multiline mode, and a match for circumflex is possible
917     when the \fIstartoffset\fP argument of \fBpcre_exec()\fP is non-zero. The
918     PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
919     .P
920 nigel 75 Note that the sequences \eA, \eZ, and \ez can be used to match the start and
921 nigel 63 end of the subject in both modes, and if all branches of a pattern start with
922 nigel 91 \eA it is always anchored, whether or not PCRE_MULTILINE is set.
923 nigel 75 .
924     .
925 ph10 514 .\" HTML <a name="fullstopdot"></a>
926     .SH "FULL STOP (PERIOD, DOT) AND \eN"
927 nigel 63 .rs
928     .sp
929     Outside a character class, a dot in the pattern matches any one character in
930 nigel 91 the subject string except (by default) a character that signifies the end of a
931 nigel 93 line. In UTF-8 mode, the matched character may be more than one byte long.
932 nigel 91 .P
933 nigel 93 When a line ending is defined as a single character, dot never matches that
934     character; when the two-character sequence CRLF is used, dot does not match CR
935     if it is immediately followed by LF, but otherwise it matches all characters
936     (including isolated CRs and LFs). When any Unicode line endings are being
937     recognized, dot does not match CR or LF or any of the other line ending
938     characters.
939     .P
940 nigel 91 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
941 nigel 93 option is set, a dot matches any one character, without exception. If the
942     two-character sequence CRLF is present in the subject string, it takes two dots
943     to match it.
944 nigel 91 .P
945     The handling of dot is entirely independent of the handling of circumflex and
946     dollar, the only relationship being that they both involve newlines. Dot has no
947     special meaning in a character class.
948 ph10 514 .P
949     The escape sequence \eN always behaves as a dot does when PCRE_DOTALL is not
950     set. In other words, it matches any one character except one that signifies the
951     end of a line.
952 nigel 75 .
953     .
954     .SH "MATCHING A SINGLE BYTE"
955 nigel 63 .rs
956     .sp
957 nigel 75 Outside a character class, the escape sequence \eC matches any one byte, both
958 nigel 93 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
959     characters. The feature is provided in Perl in order to match individual bytes
960     in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes,
961     what remains in the string may be a malformed UTF-8 string. For this reason,
962     the \eC escape sequence is best avoided.
963 nigel 75 .P
964     PCRE does not allow \eC to appear in lookbehind assertions
965     .\" HTML <a href="#lookbehind">
966     .\" </a>
967     (described below),
968     .\"
969     because in UTF-8 mode this would make it impossible to calculate the length of
970     the lookbehind.
971     .
972     .
973     .\" HTML <a name="characterclass"></a>
974     .SH "SQUARE BRACKETS AND CHARACTER CLASSES"
975 nigel 63 .rs
976     .sp
977     An opening square bracket introduces a character class, terminated by a closing
978 ph10 461 square bracket. A closing square bracket on its own is not special by default.
979     However, if the PCRE_JAVASCRIPT_COMPAT option is set, a lone closing square
980 ph10 456 bracket causes a compile-time error. If a closing square bracket is required as
981     a member of the class, it should be the first data character in the class
982     (after an initial circumflex, if present) or escaped with a backslash.
983 nigel 75 .P
984 nigel 63 A character class matches a single character in the subject. In UTF-8 mode, the
985 ph10 456 character may be more than one byte long. A matched character must be in the
986     set of characters defined by the class, unless the first character in the class
987 nigel 63 definition is a circumflex, in which case the subject character must not be in
988     the set defined by the class. If a circumflex is actually required as a member
989     of the class, ensure it is not the first character, or escape it with a
990     backslash.
991 nigel 75 .P
992 nigel 63 For example, the character class [aeiou] matches any lower case vowel, while
993     [^aeiou] matches any character that is not a lower case vowel. Note that a
994 nigel 75 circumflex is just a convenient notation for specifying the characters that
995     are in the class by enumerating those that are not. A class that starts with a
996 ph10 456 circumflex is not an assertion; it still consumes a character from the subject
997 nigel 75 string, and therefore it fails if the current pointer is at the end of the
998     string.
999     .P
1000 nigel 63 In UTF-8 mode, characters with values greater than 255 can be included in a
1001 nigel 75 class as a literal string of bytes, or by using the \ex{ escaping mechanism.
1002     .P
1003 nigel 63 When caseless matching is set, any letters in a class represent both their
1004     upper case and lower case versions, so for example, a caseless [aeiou] matches
1005     "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
1006 nigel 77 caseful version would. In UTF-8 mode, PCRE always understands the concept of
1007     case for characters whose values are less than 128, so caseless matching is
1008     always possible. For characters with higher values, the concept of case is
1009     supported if PCRE is compiled with Unicode property support, but not otherwise.
1010 ph10 456 If you want to use caseless matching in UTF8-mode for characters 128 and above,
1011     you must ensure that PCRE is compiled with Unicode property support as well as
1012     with UTF-8 support.
1013 nigel 75 .P
1014 nigel 93 Characters that might indicate line breaks are never treated in any special way
1015     when matching character classes, whatever line-ending sequence is in use, and
1016     whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
1017     such as [^a] always matches one of these characters.
1018 nigel 75 .P
1019 nigel 63 The minus (hyphen) character can be used to specify a range of characters in a
1020     character class. For example, [d-m] matches any letter between d and m,
1021     inclusive. If a minus character is required in a class, it must be escaped with
1022     a backslash or appear in a position where it cannot be interpreted as
1023     indicating a range, typically as the first or last character in the class.
1024 nigel 75 .P
1025 nigel 63 It is not possible to have the literal character "]" as the end character of a
1026     range. A pattern such as [W-]46] is interpreted as a class of two characters
1027     ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
1028     "-46]". However, if the "]" is escaped with a backslash it is interpreted as
1029 nigel 75 the end of range, so [W-\e]46] is interpreted as a class containing a range
1030     followed by two other characters. The octal or hexadecimal representation of
1031     "]" can also be used to end a range.
1032     .P
1033 nigel 63 Ranges operate in the collating sequence of character values. They can also be
1034 nigel 75 used for characters specified numerically, for example [\e000-\e037]. In UTF-8
1035 nigel 63 mode, ranges can include characters whose values are greater than 255, for
1036 nigel 75 example [\ex{100}-\ex{2ff}].
1037     .P
1038 nigel 63 If a range that includes letters is used when caseless matching is set, it
1039     matches the letters in either case. For example, [W-c] is equivalent to
1040 nigel 75 [][\e\e^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
1041 ph10 139 tables for a French locale are in use, [\exc8-\excb] matches accented E
1042 nigel 75 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
1043     characters with values greater than 128 only when it is compiled with Unicode
1044     property support.
1045     .P
1046 ph10 518 The character types \ed, \eD, \eh, \eH, \ep, \eP, \es, \eS, \ev, \eV, \ew, and
1047     \eW may also appear in a character class, and add the characters that they
1048     match to the class. For example, [\edABCDEF] matches any hexadecimal digit. A
1049     circumflex can conveniently be used with the upper case character types to
1050     specify a more restricted set of characters than the matching lower case type.
1051     For example, the class [^\eW_] matches any letter or digit, but not underscore.
1052 nigel 75 .P
1053     The only metacharacters that are recognized in character classes are backslash,
1054     hyphen (only where it can be interpreted as specifying a range), circumflex
1055     (only at the start), opening square bracket (only when it can be interpreted as
1056     introducing a POSIX class name - see the next section), and the terminating
1057     closing square bracket. However, escaping other non-alphanumeric characters
1058     does no harm.
1059     .
1060     .
1061     .SH "POSIX CHARACTER CLASSES"
1062 nigel 63 .rs
1063     .sp
1064 nigel 75 Perl supports the POSIX notation for character classes. This uses names
1065 nigel 63 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
1066     this notation. For example,
1067 nigel 75 .sp
1068 nigel 63 [01[:alpha:]%]
1069 nigel 75 .sp
1070 nigel 63 matches "0", "1", any alphabetic character, or "%". The supported class names
1071 ph10 518 are:
1072 nigel 75 .sp
1073 nigel 63 alnum letters and digits
1074     alpha letters
1075     ascii character codes 0 - 127
1076     blank space or tab only
1077     cntrl control characters
1078 nigel 75 digit decimal digits (same as \ed)
1079 nigel 63 graph printing characters, excluding space
1080     lower lower case letters
1081     print printing characters, including space
1082 ph10 518 punct printing characters, excluding letters and digits and space
1083 nigel 75 space white space (not quite the same as \es)
1084 nigel 63 upper upper case letters
1085 nigel 75 word "word" characters (same as \ew)
1086 nigel 63 xdigit hexadecimal digits
1087 nigel 75 .sp
1088 nigel 63 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
1089     space (32). Notice that this list includes the VT character (code 11). This
1090 nigel 75 makes "space" different to \es, which does not include VT (for Perl
1091 nigel 63 compatibility).
1092 nigel 75 .P
1093 nigel 63 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
1094     5.8. Another Perl extension is negation, which is indicated by a ^ character
1095     after the colon. For example,
1096 nigel 75 .sp
1097 nigel 63 [12[:^digit:]]
1098 nigel 75 .sp
1099 nigel 63 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
1100     syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
1101     supported, and an error is given if they are encountered.
1102 nigel 75 .P
1103 ph10 518 By default, in UTF-8 mode, characters with values greater than 128 do not match
1104     any of the POSIX character classes. However, if the PCRE_UCP option is passed
1105     to \fBpcre_compile()\fP, some of the classes are changed so that Unicode
1106     character properties are used. This is achieved by replacing the POSIX classes
1107     by other sequences, as follows:
1108     .sp
1109     [:alnum:] becomes \ep{Xan}
1110     [:alpha:] becomes \ep{L}
1111     [:blank:] becomes \eh
1112     [:digit:] becomes \ep{Nd}
1113     [:lower:] becomes \ep{Ll}
1114     [:space:] becomes \ep{Xps}
1115     [:upper:] becomes \ep{Lu}
1116     [:word:] becomes \ep{Xwd}
1117     .sp
1118     Negated versions, such as [:^alpha:] use \eP instead of \ep. The other POSIX
1119     classes are unchanged, and match only characters with code points less than
1120     128.
1121 nigel 75 .
1122     .
1123     .SH "VERTICAL BAR"
1124 nigel 63 .rs
1125     .sp
1126     Vertical bar characters are used to separate alternative patterns. For example,
1127     the pattern
1128 nigel 75 .sp
1129 nigel 63 gilbert|sullivan
1130 nigel 75 .sp
1131 nigel 63 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
1132 nigel 91 and an empty alternative is permitted (matching the empty string). The matching
1133     process tries each alternative in turn, from left to right, and the first one
1134     that succeeds is used. If the alternatives are within a subpattern
1135 nigel 75 .\" HTML <a href="#subpattern">
1136     .\" </a>
1137     (defined below),
1138     .\"
1139     "succeeds" means matching the rest of the main pattern as well as the
1140     alternative in the subpattern.
1141     .
1142     .
1143     .SH "INTERNAL OPTION SETTING"
1144 nigel 63 .rs
1145     .sp
1146     The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1147 ph10 231 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1148     the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1149     The option letters are
1150 nigel 75 .sp
1151 nigel 63 i for PCRE_CASELESS
1152     m for PCRE_MULTILINE
1153     s for PCRE_DOTALL
1154     x for PCRE_EXTENDED
1155 nigel 75 .sp
1156 nigel 63 For example, (?im) sets caseless, multiline matching. It is also possible to
1157     unset these options by preceding the letter with a hyphen, and a combined
1158     setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1159     PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1160     permitted. If a letter appears both before and after the hyphen, the option is
1161     unset.
1162 nigel 75 .P
1163 ph10 231 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1164     changed in the same way as the Perl-compatible options by using the characters
1165     J, U and X respectively.
1166     .P
1167 ph10 412 When one of these option changes occurs at top level (that is, not inside
1168     subpattern parentheses), the change applies to the remainder of the pattern
1169     that follows. If the change is placed right at the start of a pattern, PCRE
1170     extracts it into the global options (and it will therefore show up in data
1171     extracted by the \fBpcre_fullinfo()\fP function).
1172 nigel 75 .P
1173 nigel 93 An option change within a subpattern (see below for a description of
1174     subpatterns) affects only that part of the current pattern that follows it, so
1175 nigel 75 .sp
1176 nigel 63 (a(?i)b)c
1177 nigel 75 .sp
1178 nigel 63 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1179     By this means, options can be made to have different settings in different
1180     parts of the pattern. Any changes made in one alternative do carry on
1181     into subsequent branches within the same subpattern. For example,
1182 nigel 75 .sp
1183 nigel 63 (a(?i)b|c)
1184 nigel 75 .sp
1185 nigel 63 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1186     branch is abandoned before the option setting. This is because the effects of
1187     option settings happen at compile time. There would be some very weird
1188     behaviour otherwise.
1189 ph10 251 .P
1190     \fBNote:\fP There are other PCRE-specific options that can be set by the
1191     application when the compile or match functions are called. In some cases the
1192 ph10 412 pattern can contain special leading sequences such as (*CRLF) to override what
1193     the application has set or what has been defaulted. Details are given in the
1194     section entitled
1195 ph10 251 .\" HTML <a href="#newlineseq">
1196     .\" </a>
1197     "Newline sequences"
1198     .\"
1199 ph10 518 above. There are also the (*UTF8) and (*UCP) leading sequences that can be used
1200     to set UTF-8 and Unicode property modes; they are equivalent to setting the
1201     PCRE_UTF8 and the PCRE_UCP options, respectively.
1202 nigel 75 .
1203     .
1204     .\" HTML <a name="subpattern"></a>
1205 nigel 63 .SH SUBPATTERNS
1206     .rs
1207     .sp
1208     Subpatterns are delimited by parentheses (round brackets), which can be nested.
1209 nigel 75 Turning part of a pattern into a subpattern does two things:
1210     .sp
1211 nigel 63 1. It localizes a set of alternatives. For example, the pattern
1212 nigel 75 .sp
1213 nigel 63 cat(aract|erpillar|)
1214 nigel 75 .sp
1215 nigel 63 matches one of the words "cat", "cataract", or "caterpillar". Without the
1216 nigel 93 parentheses, it would match "cataract", "erpillar" or an empty string.
1217 nigel 75 .sp
1218     2. It sets up the subpattern as a capturing subpattern. This means that, when
1219     the whole pattern matches, that portion of the subject string that matched the
1220     subpattern is passed back to the caller via the \fIovector\fP argument of
1221     \fBpcre_exec()\fP. Opening parentheses are counted from left to right (starting
1222     from 1) to obtain numbers for the capturing subpatterns.
1223     .P
1224 nigel 63 For example, if the string "the red king" is matched against the pattern
1225 nigel 75 .sp
1226 nigel 63 the ((red|white) (king|queen))
1227 nigel 75 .sp
1228 nigel 63 the captured substrings are "red king", "red", and "king", and are numbered 1,
1229     2, and 3, respectively.
1230 nigel 75 .P
1231 nigel 63 The fact that plain parentheses fulfil two functions is not always helpful.
1232     There are often times when a grouping subpattern is required without a
1233     capturing requirement. If an opening parenthesis is followed by a question mark
1234     and a colon, the subpattern does not do any capturing, and is not counted when
1235     computing the number of any subsequent capturing subpatterns. For example, if
1236     the string "the white queen" is matched against the pattern
1237 nigel 75 .sp
1238 nigel 63 the ((?:red|white) (king|queen))
1239 nigel 75 .sp
1240 nigel 63 the captured substrings are "white queen" and "queen", and are numbered 1 and
1241 nigel 93 2. The maximum number of capturing subpatterns is 65535.
1242 nigel 75 .P
1243 nigel 63 As a convenient shorthand, if any option settings are required at the start of
1244     a non-capturing subpattern, the option letters may appear between the "?" and
1245     the ":". Thus the two patterns
1246 nigel 75 .sp
1247 nigel 63 (?i:saturday|sunday)
1248     (?:(?i)saturday|sunday)
1249 nigel 75 .sp
1250 nigel 63 match exactly the same set of strings. Because alternative branches are tried
1251     from left to right, and options are not reset until the end of the subpattern
1252     is reached, an option setting in one branch does affect subsequent branches, so
1253     the above patterns match "SUNDAY" as well as "Saturday".
1254 nigel 75 .
1255     .
1256 ph10 456 .\" HTML <a name="dupsubpatternnumber"></a>
1257 ph10 175 .SH "DUPLICATE SUBPATTERN NUMBERS"
1258     .rs
1259     .sp
1260 ph10 182 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1261     the same numbers for its capturing parentheses. Such a subpattern starts with
1262     (?| and is itself a non-capturing subpattern. For example, consider this
1263 ph10 175 pattern:
1264     .sp
1265     (?|(Sat)ur|(Sun))day
1266 ph10 182 .sp
1267     Because the two alternatives are inside a (?| group, both sets of capturing
1268     parentheses are numbered one. Thus, when the pattern matches, you can look
1269     at captured substring number one, whichever alternative matched. This construct
1270     is useful when you want to capture part, but not all, of one of a number of
1271     alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1272 ph10 175 number is reset at the start of each branch. The numbers of any capturing
1273 ph10 182 buffers that follow the subpattern start after the highest number used in any
1274     branch. The following example is taken from the Perl documentation.
1275 ph10 175 The numbers underneath show in which buffer the captured content will be
1276     stored.
1277     .sp
1278     # before ---------------branch-reset----------- after
1279     / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1280     # 1 2 2 3 2 3 4
1281 ph10 182 .sp
1282 ph10 488 A back reference to a numbered subpattern uses the most recent value that is
1283     set for that number by any subpattern. The following pattern matches "abcabc"
1284     or "defdef":
1285 ph10 456 .sp
1286 ph10 461 /(?|(abc)|(def))\e1/
1287 ph10 456 .sp
1288     In contrast, a recursive or "subroutine" call to a numbered subpattern always
1289 ph10 461 refers to the first one in the pattern with the given number. The following
1290 ph10 456 pattern matches "abcabc" or "defabc":
1291     .sp
1292     /(?|(abc)|(def))(?1)/
1293     .sp
1294 ph10 459 If a
1295     .\" HTML <a href="#conditions">
1296     .\" </a>
1297     condition test
1298     .\"
1299     for a subpattern's having matched refers to a non-unique number, the test is
1300     true if any of the subpatterns of that number have matched.
1301     .P
1302     An alternative approach to using this "branch reset" feature is to use
1303 ph10 175 duplicate named subpatterns, as described in the next section.
1304     .
1305     .
1306 nigel 75 .SH "NAMED SUBPATTERNS"
1307 nigel 63 .rs
1308     .sp
1309     Identifying capturing parentheses by number is simple, but it can be very hard
1310     to keep track of the numbers in complicated regular expressions. Furthermore,
1311 nigel 75 if an expression is modified, the numbers may change. To help with this
1312 nigel 93 difficulty, PCRE supports the naming of subpatterns. This feature was not
1313     added to Perl until release 5.10. Python had the feature earlier, and PCRE
1314     introduced it at release 4.0, using the Python syntax. PCRE now supports both
1315 ph10 459 the Perl and the Python syntax. Perl allows identically numbered subpatterns to
1316     have different names, but PCRE does not.
1317 nigel 93 .P
1318     In PCRE, a subpattern can be named in one of three ways: (?<name>...) or
1319     (?'name'...) as in Perl, or (?P<name>...) as in Python. References to capturing
1320 nigel 91 parentheses from other parts of the pattern, such as
1321     .\" HTML <a href="#backreferences">
1322     .\" </a>
1323 ph10 488 back references,
1324 nigel 91 .\"
1325     .\" HTML <a href="#recursion">
1326     .\" </a>
1327     recursion,
1328     .\"
1329     and
1330     .\" HTML <a href="#conditions">
1331     .\" </a>
1332     conditions,
1333     .\"
1334     can be made by name as well as by number.
1335 nigel 75 .P
1336 nigel 91 Names consist of up to 32 alphanumeric characters and underscores. Named
1337 nigel 93 capturing parentheses are still allocated numbers as well as names, exactly as
1338     if the names were not present. The PCRE API provides function calls for
1339     extracting the name-to-number translation table from a compiled pattern. There
1340     is also a convenience function for extracting a captured substring by name.
1341 nigel 91 .P
1342     By default, a name must be unique within a pattern, but it is possible to relax
1343 ph10 457 this constraint by setting the PCRE_DUPNAMES option at compile time. (Duplicate
1344 ph10 461 names are also always permitted for subpatterns with the same number, set up as
1345 ph10 457 described in the previous section.) Duplicate names can be useful for patterns
1346     where only one instance of the named parentheses can match. Suppose you want to
1347     match the name of a weekday, either as a 3-letter abbreviation or as the full
1348     name, and in both cases you want to extract the abbreviation. This pattern
1349     (ignoring the line breaks) does the job:
1350 nigel 91 .sp
1351 nigel 93 (?<DN>Mon|Fri|Sun)(?:day)?|
1352     (?<DN>Tue)(?:sday)?|
1353     (?<DN>Wed)(?:nesday)?|
1354     (?<DN>Thu)(?:rsday)?|
1355     (?<DN>Sat)(?:urday)?
1356 nigel 91 .sp
1357     There are five capturing substrings, but only one is ever set after a match.
1358 ph10 182 (An alternative way of solving this problem is to use a "branch reset"
1359 ph10 175 subpattern, as described in the previous section.)
1360     .P
1361 nigel 91 The convenience function for extracting the data by name returns the substring
1362 nigel 93 for the first (and in this example, the only) subpattern of that name that
1363 ph10 461 matched. This saves searching to find which numbered subpattern it was.
1364 ph10 459 .P
1365 ph10 488 If you make a back reference to a non-unique named subpattern from elsewhere in
1366 ph10 459 the pattern, the one that corresponds to the first occurrence of the name is
1367     used. In the absence of duplicate numbers (see the previous section) this is
1368     the one with the lowest number. If you use a named reference in a condition
1369     test (see the
1370     .\"
1371     .\" HTML <a href="#conditions">
1372     .\" </a>
1373     section about conditions
1374     .\"
1375 ph10 461 below), either to check whether a subpattern has matched, or to check for
1376 ph10 459 recursion, all subpatterns with the same name are tested. If the condition is
1377     true for any one of them, the overall condition is true. This is the same
1378     behaviour as testing by number. For further details of the interfaces for
1379     handling named subpatterns, see the
1380 nigel 63 .\" HREF
1381 nigel 75 \fBpcreapi\fP
1382 nigel 63 .\"
1383     documentation.
1384 ph10 385 .P
1385     \fBWarning:\fP You cannot use different names to distinguish between two
1386 ph10 457 subpatterns with the same number because PCRE uses only the numbers when
1387     matching. For this reason, an error is given at compile time if different names
1388     are given to subpatterns with the same number. However, you can give the same
1389     name to subpatterns with the same number, even when PCRE_DUPNAMES is not set.
1390 nigel 75 .
1391     .
1392 nigel 63 .SH REPETITION
1393     .rs
1394     .sp
1395     Repetition is specified by quantifiers, which can follow any of the following
1396     items:
1397 nigel 75 .sp
1398 nigel 63 a literal data character
1399 nigel 93 the dot metacharacter
1400 nigel 75 the \eC escape sequence
1401     the \eX escape sequence (in UTF-8 mode with Unicode properties)
1402 nigel 93 the \eR escape sequence
1403 nigel 75 an escape such as \ed that matches a single character
1404 nigel 63 a character class
1405     a back reference (see next section)
1406     a parenthesized subpattern (unless it is an assertion)
1407 ph10 461 a recursive or "subroutine" call to a subpattern
1408 nigel 75 .sp
1409 nigel 63 The general repetition quantifier specifies a minimum and maximum number of
1410     permitted matches, by giving the two numbers in curly brackets (braces),
1411     separated by a comma. The numbers must be less than 65536, and the first must
1412     be less than or equal to the second. For example:
1413 nigel 75 .sp
1414 nigel 63 z{2,4}
1415 nigel 75 .sp
1416 nigel 63 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1417     character. If the second number is omitted, but the comma is present, there is
1418     no upper limit; if the second number and the comma are both omitted, the
1419     quantifier specifies an exact number of required matches. Thus
1420 nigel 75 .sp
1421 nigel 63 [aeiou]{3,}
1422 nigel 75 .sp
1423 nigel 63 matches at least 3 successive vowels, but may match many more, while
1424 nigel 75 .sp
1425     \ed{8}
1426     .sp
1427 nigel 63 matches exactly 8 digits. An opening curly bracket that appears in a position
1428     where a quantifier is not allowed, or one that does not match the syntax of a
1429     quantifier, is taken as a literal character. For example, {,6} is not a
1430     quantifier, but a literal string of four characters.
1431 nigel 75 .P
1432 nigel 63 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1433 nigel 75 bytes. Thus, for example, \ex{100}{2} matches two UTF-8 characters, each of
1434     which is represented by a two-byte sequence. Similarly, when Unicode property
1435     support is available, \eX{3} matches three Unicode extended sequences, each of
1436     which may be several bytes long (and they may be of different lengths).
1437     .P
1438 nigel 63 The quantifier {0} is permitted, causing the expression to behave as if the
1439 ph10 345 previous item and the quantifier were not present. This may be useful for
1440     subpatterns that are referenced as
1441 ph10 335 .\" HTML <a href="#subpatternsassubroutines">
1442     .\" </a>
1443     subroutines
1444     .\"
1445 ph10 345 from elsewhere in the pattern. Items other than subpatterns that have a {0}
1446 ph10 335 quantifier are omitted from the compiled pattern.
1447 nigel 75 .P
1448 nigel 93 For convenience, the three most common quantifiers have single-character
1449     abbreviations:
1450 nigel 75 .sp
1451 nigel 63 * is equivalent to {0,}
1452     + is equivalent to {1,}
1453     ? is equivalent to {0,1}
1454 nigel 75 .sp
1455 nigel 63 It is possible to construct infinite loops by following a subpattern that can
1456     match no characters with a quantifier that has no upper limit, for example:
1457 nigel 75 .sp
1458 nigel 63 (a?)*
1459 nigel 75 .sp
1460 nigel 63 Earlier versions of Perl and PCRE used to give an error at compile time for
1461     such patterns. However, because there are cases where this can be useful, such
1462     patterns are now accepted, but if any repetition of the subpattern does in fact
1463     match no characters, the loop is forcibly broken.
1464 nigel 75 .P
1465 nigel 63 By default, the quantifiers are "greedy", that is, they match as much as
1466     possible (up to the maximum number of permitted times), without causing the
1467     rest of the pattern to fail. The classic example of where this gives problems
1468 nigel 75 is in trying to match comments in C programs. These appear between /* and */
1469     and within the comment, individual * and / characters may appear. An attempt to
1470     match C comments by applying the pattern
1471     .sp
1472     /\e*.*\e*/
1473     .sp
1474 nigel 63 to the string
1475 nigel 75 .sp
1476     /* first comment */ not comment /* second comment */
1477     .sp
1478 nigel 63 fails, because it matches the entire string owing to the greediness of the .*
1479     item.
1480 nigel 75 .P
1481 nigel 63 However, if a quantifier is followed by a question mark, it ceases to be
1482     greedy, and instead matches the minimum number of times possible, so the
1483     pattern
1484 nigel 75 .sp
1485     /\e*.*?\e*/
1486     .sp
1487 nigel 63 does the right thing with the C comments. The meaning of the various
1488     quantifiers is not otherwise changed, just the preferred number of matches.
1489     Do not confuse this use of question mark with its use as a quantifier in its
1490     own right. Because it has two uses, it can sometimes appear doubled, as in
1491 nigel 75 .sp
1492     \ed??\ed
1493     .sp
1494 nigel 63 which matches one digit by preference, but can match two if that is the only
1495     way the rest of the pattern matches.
1496 nigel 75 .P
1497 nigel 93 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1498 nigel 63 the quantifiers are not greedy by default, but individual ones can be made
1499     greedy by following them with a question mark. In other words, it inverts the
1500     default behaviour.
1501 nigel 75 .P
1502 nigel 63 When a parenthesized subpattern is quantified with a minimum repeat count that
1503 nigel 75 is greater than 1 or with a limited maximum, more memory is required for the
1504 nigel 63 compiled pattern, in proportion to the size of the minimum or maximum.
1505 nigel 75 .P
1506 nigel 63 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1507 nigel 93 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1508 nigel 63 implicitly anchored, because whatever follows will be tried against every
1509     character position in the subject string, so there is no point in retrying the
1510     overall match at any position after the first. PCRE normally treats such a
1511 nigel 75 pattern as though it were preceded by \eA.
1512     .P
1513 nigel 63 In cases where it is known that the subject string contains no newlines, it is
1514     worth setting PCRE_DOTALL in order to obtain this optimization, or
1515     alternatively using ^ to indicate anchoring explicitly.
1516 nigel 75 .P
1517 nigel 63 However, there is one situation where the optimization cannot be used. When .*
1518 ph10 488 is inside capturing parentheses that are the subject of a back reference
1519 nigel 93 elsewhere in the pattern, a match at the start may fail where a later one
1520     succeeds. Consider, for example:
1521 nigel 75 .sp
1522     (.*)abc\e1
1523     .sp
1524 nigel 63 If the subject is "xyz123abc123" the match point is the fourth character. For
1525     this reason, such a pattern is not implicitly anchored.
1526 nigel 75 .P
1527 nigel 63 When a capturing subpattern is repeated, the value captured is the substring
1528     that matched the final iteration. For example, after
1529 nigel 75 .sp
1530     (tweedle[dume]{3}\es*)+
1531     .sp
1532 nigel 63 has matched "tweedledum tweedledee" the value of the captured substring is
1533     "tweedledee". However, if there are nested capturing subpatterns, the
1534     corresponding captured values may have been set in previous iterations. For
1535     example, after
1536 nigel 75 .sp
1537 nigel 63 /(a|(b))+/
1538 nigel 75 .sp
1539 nigel 63 matches "aba" the value of the second captured substring is "b".
1540 nigel 75 .
1541     .
1542     .\" HTML <a name="atomicgroup"></a>
1543     .SH "ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS"
1544 nigel 63 .rs
1545     .sp
1546 nigel 93 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1547     repetition, failure of what follows normally causes the repeated item to be
1548     re-evaluated to see if a different number of repeats allows the rest of the
1549     pattern to match. Sometimes it is useful to prevent this, either to change the
1550     nature of the match, or to cause it fail earlier than it otherwise might, when
1551     the author of the pattern knows there is no point in carrying on.
1552 nigel 75 .P
1553     Consider, for example, the pattern \ed+foo when applied to the subject line
1554     .sp
1555 nigel 63 123456bar
1556 nigel 75 .sp
1557 nigel 63 After matching all 6 digits and then failing to match "foo", the normal
1558 nigel 75 action of the matcher is to try again with only 5 digits matching the \ed+
1559 nigel 63 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1560     (a term taken from Jeffrey Friedl's book) provides the means for specifying
1561     that once a subpattern has matched, it is not to be re-evaluated in this way.
1562 nigel 75 .P
1563 nigel 93 If we use atomic grouping for the previous example, the matcher gives up
1564 nigel 63 immediately on failing to match "foo" the first time. The notation is a kind of
1565     special parenthesis, starting with (?> as in this example:
1566 nigel 75 .sp
1567     (?>\ed+)foo
1568     .sp
1569 nigel 63 This kind of parenthesis "locks up" the part of the pattern it contains once
1570     it has matched, and a failure further into the pattern is prevented from
1571     backtracking into it. Backtracking past it to previous items, however, works as
1572     normal.
1573 nigel 75 .P
1574 nigel 63 An alternative description is that a subpattern of this type matches the string
1575     of characters that an identical standalone pattern would match, if anchored at
1576     the current point in the subject string.
1577 nigel 75 .P
1578 nigel 63 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1579     the above example can be thought of as a maximizing repeat that must swallow
1580 nigel 75 everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1581 nigel 63 number of digits they match in order to make the rest of the pattern match,
1582 nigel 75 (?>\ed+) can only match an entire sequence of digits.
1583     .P
1584 nigel 63 Atomic groups in general can of course contain arbitrarily complicated
1585     subpatterns, and can be nested. However, when the subpattern for an atomic
1586     group is just a single repeated item, as in the example above, a simpler
1587     notation, called a "possessive quantifier" can be used. This consists of an
1588     additional + character following a quantifier. Using this notation, the
1589     previous example can be rewritten as
1590 nigel 75 .sp
1591     \ed++foo
1592     .sp
1593 ph10 208 Note that a possessive quantifier can be used with an entire group, for
1594     example:
1595     .sp
1596     (abc|xyz){2,3}+
1597     .sp
1598 nigel 63 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1599     option is ignored. They are a convenient notation for the simpler forms of
1600 nigel 93 atomic group. However, there is no difference in the meaning of a possessive
1601     quantifier and the equivalent atomic group, though there may be a performance
1602     difference; possessive quantifiers should be slightly faster.
1603 nigel 75 .P
1604 nigel 93 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1605     Jeffrey Friedl originated the idea (and the name) in the first edition of his
1606     book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1607     package, and PCRE copied it from there. It ultimately found its way into Perl
1608     at release 5.10.
1609 nigel 75 .P
1610 nigel 93 PCRE has an optimization that automatically "possessifies" certain simple
1611     pattern constructs. For example, the sequence A+B is treated as A++B because
1612     there is no point in backtracking into a sequence of A's when B must follow.
1613     .P
1614 nigel 63 When a pattern contains an unlimited repeat inside a subpattern that can itself
1615     be repeated an unlimited number of times, the use of an atomic group is the
1616     only way to avoid some failing matches taking a very long time indeed. The
1617     pattern
1618 nigel 75 .sp
1619     (\eD+|<\ed+>)*[!?]
1620     .sp
1621 nigel 63 matches an unlimited number of substrings that either consist of non-digits, or
1622     digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1623     quickly. However, if it is applied to
1624 nigel 75 .sp
1625 nigel 63 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1626 nigel 75 .sp
1627 nigel 63 it takes a long time before reporting failure. This is because the string can
1628 nigel 75 be divided between the internal \eD+ repeat and the external * repeat in a
1629     large number of ways, and all have to be tried. (The example uses [!?] rather
1630     than a single character at the end, because both PCRE and Perl have an
1631     optimization that allows for fast failure when a single character is used. They
1632     remember the last single character that is required for a match, and fail early
1633     if it is not present in the string.) If the pattern is changed so that it uses
1634     an atomic group, like this:
1635     .sp
1636     ((?>\eD+)|<\ed+>)*[!?]
1637     .sp
1638 nigel 63 sequences of non-digits cannot be broken, and failure happens quickly.
1639 nigel 75 .
1640     .
1641     .\" HTML <a name="backreferences"></a>
1642     .SH "BACK REFERENCES"
1643 nigel 63 .rs
1644     .sp
1645     Outside a character class, a backslash followed by a digit greater than 0 (and
1646     possibly further digits) is a back reference to a capturing subpattern earlier
1647     (that is, to its left) in the pattern, provided there have been that many
1648     previous capturing left parentheses.
1649 nigel 75 .P
1650 nigel 63 However, if the decimal number following the backslash is less than 10, it is
1651     always taken as a back reference, and causes an error only if there are not
1652     that many capturing left parentheses in the entire pattern. In other words, the
1653     parentheses that are referenced need not be to the left of the reference for
1654 nigel 91 numbers less than 10. A "forward back reference" of this type can make sense
1655     when a repetition is involved and the subpattern to the right has participated
1656     in an earlier iteration.
1657     .P
1658 nigel 93 It is not possible to have a numerical "forward back reference" to a subpattern
1659     whose number is 10 or more using this syntax because a sequence such as \e50 is
1660     interpreted as a character defined in octal. See the subsection entitled
1661 nigel 91 "Non-printing characters"
1662 nigel 75 .\" HTML <a href="#digitsafterbackslash">
1663     .\" </a>
1664     above
1665     .\"
1666 nigel 93 for further details of the handling of digits following a backslash. There is
1667     no such problem when named parentheses are used. A back reference to any
1668     subpattern is possible using named parentheses (see below).
1669 nigel 75 .P
1670 nigel 93 Another way of avoiding the ambiguity inherent in the use of digits following a
1671     backslash is to use the \eg escape sequence, which is a feature introduced in
1672 ph10 208 Perl 5.10. This escape must be followed by an unsigned number or a negative
1673     number, optionally enclosed in braces. These examples are all identical:
1674 nigel 93 .sp
1675     (ring), \e1
1676     (ring), \eg1
1677     (ring), \eg{1}
1678     .sp
1679 ph10 208 An unsigned number specifies an absolute reference without the ambiguity that
1680     is present in the older syntax. It is also useful when literal digits follow
1681     the reference. A negative number is a relative reference. Consider this
1682     example:
1683 nigel 93 .sp
1684     (abc(def)ghi)\eg{-1}
1685     .sp
1686     The sequence \eg{-1} is a reference to the most recently started capturing
1687     subpattern before \eg, that is, is it equivalent to \e2. Similarly, \eg{-2}
1688     would be equivalent to \e1. The use of relative references can be helpful in
1689     long patterns, and also in patterns that are created by joining together
1690     fragments that contain references within themselves.
1691     .P
1692 nigel 63 A back reference matches whatever actually matched the capturing subpattern in
1693     the current subject string, rather than anything matching the subpattern
1694     itself (see
1695     .\" HTML <a href="#subpatternsassubroutines">
1696     .\" </a>
1697     "Subpatterns as subroutines"
1698     .\"
1699     below for a way of doing that). So the pattern
1700 nigel 75 .sp
1701     (sens|respons)e and \e1ibility
1702     .sp
1703 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1704     "sense and responsibility". If caseful matching is in force at the time of the
1705     back reference, the case of letters is relevant. For example,
1706 nigel 75 .sp
1707     ((?i)rah)\es+\e1
1708     .sp
1709 nigel 63 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1710     capturing subpattern is matched caselessly.
1711 nigel 75 .P
1712 ph10 171 There are several different ways of writing back references to named
1713     subpatterns. The .NET syntax \ek{name} and the Perl syntax \ek<name> or
1714     \ek'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1715     back reference syntax, in which \eg can be used for both numeric and named
1716     references, is also supported. We could rewrite the above example in any of
1717 nigel 93 the following ways:
1718 nigel 75 .sp
1719 nigel 93 (?<p1>(?i)rah)\es+\ek<p1>
1720 ph10 171 (?'p1'(?i)rah)\es+\ek{p1}
1721 nigel 91 (?P<p1>(?i)rah)\es+(?P=p1)
1722 ph10 171 (?<p1>(?i)rah)\es+\eg{p1}
1723 nigel 75 .sp
1724 nigel 91 A subpattern that is referenced by name may appear in the pattern before or
1725     after the reference.
1726     .P
1727 nigel 63 There may be more than one back reference to the same subpattern. If a
1728     subpattern has not actually been used in a particular match, any back
1729 ph10 456 references to it always fail by default. For example, the pattern
1730 nigel 75 .sp
1731     (a|(bc))\e2
1732     .sp
1733 ph10 461 always fails if it starts to match "a" rather than "bc". However, if the
1734     PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back reference to an
1735 ph10 456 unset value matches an empty string.
1736     .P
1737     Because there may be many capturing parentheses in a pattern, all digits
1738     following a backslash are taken as part of a potential back reference number.
1739     If the pattern continues with a digit character, some delimiter must be used to
1740     terminate the back reference. If the PCRE_EXTENDED option is set, this can be
1741     whitespace. Otherwise, the \eg{ syntax or an empty comment (see
1742 nigel 75 .\" HTML <a href="#comments">
1743     .\" </a>
1744     "Comments"
1745     .\"
1746     below) can be used.
1747 ph10 488 .
1748     .SS "Recursive back references"
1749     .rs
1750     .sp
1751 nigel 63 A back reference that occurs inside the parentheses to which it refers fails
1752 nigel 75 when the subpattern is first used, so, for example, (a\e1) never matches.
1753 nigel 63 However, such references can be useful inside repeated subpatterns. For
1754     example, the pattern
1755 nigel 75 .sp
1756     (a|b\e1)+
1757     .sp
1758 nigel 63 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1759     the subpattern, the back reference matches the character string corresponding
1760     to the previous iteration. In order for this to work, the pattern must be such
1761     that the first iteration does not need to match the back reference. This can be
1762     done using alternation, as in the example above, or by a quantifier with a
1763     minimum of zero.
1764 ph10 488 .P
1765     Back references of this type cause the group that they reference to be treated
1766     as an
1767     .\" HTML <a href="#atomicgroup">
1768     .\" </a>
1769     atomic group.
1770     .\"
1771     Once the whole group has been matched, a subsequent matching failure cannot
1772     cause backtracking into the middle of the group.
1773 nigel 75 .
1774     .
1775     .\" HTML <a name="bigassertions"></a>
1776 nigel 63 .SH ASSERTIONS
1777     .rs
1778     .sp
1779     An assertion is a test on the characters following or preceding the current
1780     matching point that does not actually consume any characters. The simple
1781 nigel 75 assertions coded as \eb, \eB, \eA, \eG, \eZ, \ez, ^ and $ are described
1782     .\" HTML <a href="#smallassertions">
1783     .\" </a>
1784     above.
1785     .\"
1786     .P
1787 nigel 63 More complicated assertions are coded as subpatterns. There are two kinds:
1788     those that look ahead of the current position in the subject string, and those
1789 nigel 75 that look behind it. An assertion subpattern is matched in the normal way,
1790     except that it does not cause the current matching position to be changed.
1791     .P
1792     Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1793     because it makes no sense to assert the same thing several times. If any kind
1794     of assertion contains capturing subpatterns within it, these are counted for
1795     the purposes of numbering the capturing subpatterns in the whole pattern.
1796     However, substring capturing is carried out only for positive assertions,
1797     because it does not make sense for negative assertions.
1798     .
1799     .
1800     .SS "Lookahead assertions"
1801     .rs
1802     .sp
1803 nigel 91 Lookahead assertions start with (?= for positive assertions and (?! for
1804     negative assertions. For example,
1805 nigel 75 .sp
1806     \ew+(?=;)
1807     .sp
1808 nigel 63 matches a word followed by a semicolon, but does not include the semicolon in
1809     the match, and
1810 nigel 75 .sp
1811 nigel 63 foo(?!bar)
1812 nigel 75 .sp
1813 nigel 63 matches any occurrence of "foo" that is not followed by "bar". Note that the
1814     apparently similar pattern
1815 nigel 75 .sp
1816 nigel 63 (?!foo)bar
1817 nigel 75 .sp
1818 nigel 63 does not find an occurrence of "bar" that is preceded by something other than
1819     "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1820     (?!foo) is always true when the next three characters are "bar". A
1821 nigel 75 lookbehind assertion is needed to achieve the other effect.
1822     .P
1823 nigel 63 If you want to force a matching failure at some point in a pattern, the most
1824     convenient way to do it is with (?!) because an empty string always matches, so
1825     an assertion that requires there not to be an empty string must always fail.
1826 ph10 456 The Perl 5.10 backtracking control verb (*FAIL) or (*F) is essentially a
1827     synonym for (?!).
1828 nigel 75 .
1829     .
1830     .\" HTML <a name="lookbehind"></a>
1831     .SS "Lookbehind assertions"
1832     .rs
1833     .sp
1834 nigel 63 Lookbehind assertions start with (?<= for positive assertions and (?<! for
1835     negative assertions. For example,
1836 nigel 75 .sp
1837 nigel 63 (?<!foo)bar
1838 nigel 75 .sp
1839 nigel 63 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1840     a lookbehind assertion are restricted such that all the strings it matches must
1841 nigel 91 have a fixed length. However, if there are several top-level alternatives, they
1842     do not all have to have the same fixed length. Thus
1843 nigel 75 .sp
1844 nigel 63 (?<=bullock|donkey)
1845 nigel 75 .sp
1846 nigel 63 is permitted, but
1847 nigel 75 .sp
1848 nigel 63 (?<!dogs?|cats?)
1849 nigel 75 .sp
1850 nigel 63 causes an error at compile time. Branches that match different length strings
1851     are permitted only at the top level of a lookbehind assertion. This is an
1852 ph10 454 extension compared with Perl (5.8 and 5.10), which requires all branches to
1853 nigel 63 match the same length of string. An assertion such as
1854 nigel 75 .sp
1855 nigel 63 (?<=ab(c|de))
1856 nigel 75 .sp
1857 nigel 63 is not permitted, because its single top-level branch can match two different
1858 ph10 454 lengths, but it is acceptable to PCRE if rewritten to use two top-level
1859     branches:
1860 nigel 75 .sp
1861 nigel 63 (?<=abc|abde)
1862 nigel 75 .sp
1863 ph10 172 In some cases, the Perl 5.10 escape sequence \eK
1864 ph10 168 .\" HTML <a href="#resetmatchstart">
1865     .\" </a>
1866     (see above)
1867     .\"
1868 ph10 461 can be used instead of a lookbehind assertion to get round the fixed-length
1869 ph10 454 restriction.
1870 ph10 168 .P
1871 nigel 63 The implementation of lookbehind assertions is, for each alternative, to
1872 nigel 93 temporarily move the current position back by the fixed length and then try to
1873 nigel 63 match. If there are insufficient characters before the current position, the
1874 nigel 93 assertion fails.
1875 nigel 75 .P
1876     PCRE does not allow the \eC escape (which matches a single byte in UTF-8 mode)
1877 nigel 63 to appear in lookbehind assertions, because it makes it impossible to calculate
1878 nigel 93 the length of the lookbehind. The \eX and \eR escapes, which can match
1879     different numbers of bytes, are also not permitted.
1880 nigel 75 .P
1881 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
1882     .\" </a>
1883     "Subroutine"
1884     .\"
1885     calls (see below) such as (?2) or (?&X) are permitted in lookbehinds, as long
1886 ph10 461 as the subpattern matches a fixed-length string.
1887 ph10 454 .\" HTML <a href="#recursion">
1888     .\" </a>
1889     Recursion,
1890     .\"
1891     however, is not supported.
1892     .P
1893 nigel 93 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1894 ph10 456 specify efficient matching of fixed-length strings at the end of subject
1895     strings. Consider a simple pattern such as
1896 nigel 75 .sp
1897 nigel 63 abcd$
1898 nigel 75 .sp
1899 nigel 63 when applied to a long string that does not match. Because matching proceeds
1900     from left to right, PCRE will look for each "a" in the subject and then see if
1901     what follows matches the rest of the pattern. If the pattern is specified as
1902 nigel 75 .sp
1903 nigel 63 ^.*abcd$
1904 nigel 75 .sp
1905 nigel 63 the initial .* matches the entire string at first, but when this fails (because
1906     there is no following "a"), it backtracks to match all but the last character,
1907     then all but the last two characters, and so on. Once again the search for "a"
1908     covers the entire string, from right to left, so we are no better off. However,
1909     if the pattern is written as
1910 nigel 75 .sp
1911 nigel 63 ^.*+(?<=abcd)
1912 nigel 75 .sp
1913 nigel 93 there can be no backtracking for the .*+ item; it can match only the entire
1914 nigel 63 string. The subsequent lookbehind assertion does a single test on the last four
1915     characters. If it fails, the match fails immediately. For long strings, this
1916     approach makes a significant difference to the processing time.
1917 nigel 75 .
1918     .
1919     .SS "Using multiple assertions"
1920     .rs
1921     .sp
1922 nigel 63 Several assertions (of any sort) may occur in succession. For example,
1923 nigel 75 .sp
1924     (?<=\ed{3})(?<!999)foo
1925     .sp
1926 nigel 63 matches "foo" preceded by three digits that are not "999". Notice that each of
1927     the assertions is applied independently at the same point in the subject
1928     string. First there is a check that the previous three characters are all
1929     digits, and then there is a check that the same three characters are not "999".
1930 nigel 75 This pattern does \fInot\fP match "foo" preceded by six characters, the first
1931 nigel 63 of which are digits and the last three of which are not "999". For example, it
1932     doesn't match "123abcfoo". A pattern to do that is
1933 nigel 75 .sp
1934     (?<=\ed{3}...)(?<!999)foo
1935     .sp
1936 nigel 63 This time the first assertion looks at the preceding six characters, checking
1937     that the first three are digits, and then the second assertion checks that the
1938     preceding three characters are not "999".
1939 nigel 75 .P
1940 nigel 63 Assertions can be nested in any combination. For example,
1941 nigel 75 .sp
1942 nigel 63 (?<=(?<!foo)bar)baz
1943 nigel 75 .sp
1944 nigel 63 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1945     preceded by "foo", while
1946 nigel 75 .sp
1947     (?<=\ed{3}(?!999)...)foo
1948     .sp
1949     is another pattern that matches "foo" preceded by three digits and any three
1950 nigel 63 characters that are not "999".
1951 nigel 75 .
1952     .
1953 nigel 91 .\" HTML <a name="conditions"></a>
1954 nigel 75 .SH "CONDITIONAL SUBPATTERNS"
1955 nigel 63 .rs
1956     .sp
1957     It is possible to cause the matching process to obey a subpattern
1958     conditionally or to choose between two alternative subpatterns, depending on
1959 ph10 461 the result of an assertion, or whether a specific capturing subpattern has
1960 ph10 456 already been matched. The two possible forms of conditional subpattern are:
1961 nigel 75 .sp
1962 nigel 63 (?(condition)yes-pattern)
1963     (?(condition)yes-pattern|no-pattern)
1964 nigel 75 .sp
1965 nigel 63 If the condition is satisfied, the yes-pattern is used; otherwise the
1966     no-pattern (if present) is used. If there are more than two alternatives in the
1967     subpattern, a compile-time error occurs.
1968 nigel 75 .P
1969 nigel 93 There are four kinds of condition: references to subpatterns, references to
1970     recursion, a pseudo-condition called DEFINE, and assertions.
1971     .
1972     .SS "Checking for a used subpattern by number"
1973     .rs
1974     .sp
1975     If the text between the parentheses consists of a sequence of digits, the
1976 ph10 456 condition is true if a capturing subpattern of that number has previously
1977 ph10 461 matched. If there is more than one capturing subpattern with the same number
1978     (see the earlier
1979 ph10 456 .\"
1980     .\" HTML <a href="#recursion">
1981     .\" </a>
1982     section about duplicate subpattern numbers),
1983     .\"
1984     the condition is true if any of them have been set. An alternative notation is
1985     to precede the digits with a plus or minus sign. In this case, the subpattern
1986     number is relative rather than absolute. The most recently opened parentheses
1987     can be referenced by (?(-1), the next most recent by (?(-2), and so on. In
1988     looping constructs it can also make sense to refer to subsequent groups with
1989     constructs such as (?(+2).
1990 nigel 91 .P
1991     Consider the following pattern, which contains non-significant white space to
1992     make it more readable (assume the PCRE_EXTENDED option) and to divide it into
1993     three parts for ease of discussion:
1994 nigel 75 .sp
1995     ( \e( )? [^()]+ (?(1) \e) )
1996     .sp
1997 nigel 63 The first part matches an optional opening parenthesis, and if that
1998     character is present, sets it as the first captured substring. The second part
1999     matches one or more characters that are not parentheses. The third part is a
2000     conditional subpattern that tests whether the first set of parentheses matched
2001     or not. If they did, that is, if subject started with an opening parenthesis,
2002     the condition is true, and so the yes-pattern is executed and a closing
2003     parenthesis is required. Otherwise, since no-pattern is not present, the
2004     subpattern matches nothing. In other words, this pattern matches a sequence of
2005 nigel 93 non-parentheses, optionally enclosed in parentheses.
2006 ph10 167 .P
2007 ph10 172 If you were embedding this pattern in a larger one, you could use a relative
2008 ph10 167 reference:
2009     .sp
2010     ...other stuff... ( \e( )? [^()]+ (?(-1) \e) ) ...
2011     .sp
2012     This makes the fragment independent of the parentheses in the larger pattern.
2013 nigel 93 .
2014     .SS "Checking for a used subpattern by name"
2015     .rs
2016 nigel 91 .sp
2017 nigel 93 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a used
2018     subpattern by name. For compatibility with earlier versions of PCRE, which had
2019     this facility before Perl, the syntax (?(name)...) is also recognized. However,
2020     there is a possible ambiguity with this syntax, because subpattern names may
2021     consist entirely of digits. PCRE looks first for a named subpattern; if it
2022     cannot find one and the name consists entirely of digits, PCRE looks for a
2023     subpattern of that number, which must be greater than zero. Using subpattern
2024     names that consist entirely of digits is not recommended.
2025     .P
2026     Rewriting the above example to use a named subpattern gives this:
2027 nigel 91 .sp
2028 nigel 93 (?<OPEN> \e( )? [^()]+ (?(<OPEN>) \e) )
2029     .sp
2030 ph10 461 If the name used in a condition of this kind is a duplicate, the test is
2031     applied to all subpatterns of the same name, and is true if any one of them has
2032 ph10 459 matched.
2033 nigel 93 .
2034     .SS "Checking for pattern recursion"
2035     .rs
2036     .sp
2037 nigel 91 If the condition is the string (R), and there is no subpattern with the name R,
2038 nigel 93 the condition is true if a recursive call to the whole pattern or any
2039     subpattern has been made. If digits or a name preceded by ampersand follow the
2040     letter R, for example:
2041     .sp
2042     (?(R3)...) or (?(R&name)...)
2043     .sp
2044 ph10 456 the condition is true if the most recent recursion is into a subpattern whose
2045 nigel 93 number or name is given. This condition does not check the entire recursion
2046 ph10 461 stack. If the name used in a condition of this kind is a duplicate, the test is
2047     applied to all subpatterns of the same name, and is true if any one of them is
2048     the most recent recursion.
2049 nigel 75 .P
2050 ph10 461 At "top level", all these recursion test conditions are false.
2051 ph10 454 .\" HTML <a href="#recursion">
2052     .\" </a>
2053 ph10 459 The syntax for recursive patterns
2054 ph10 454 .\"
2055 ph10 459 is described below.
2056 nigel 93 .
2057     .SS "Defining subpatterns for use by reference only"
2058     .rs
2059     .sp
2060     If the condition is the string (DEFINE), and there is no subpattern with the
2061     name DEFINE, the condition is always false. In this case, there may be only one
2062     alternative in the subpattern. It is always skipped if control reaches this
2063     point in the pattern; the idea of DEFINE is that it can be used to define
2064 ph10 461 "subroutines" that can be referenced from elsewhere. (The use of
2065 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
2066     .\" </a>
2067     "subroutines"
2068     .\"
2069 nigel 93 is described below.) For example, a pattern to match an IPv4 address could be
2070     written like this (ignore whitespace and line breaks):
2071     .sp
2072     (?(DEFINE) (?<byte> 2[0-4]\ed | 25[0-5] | 1\ed\ed | [1-9]?\ed) )
2073     \eb (?&byte) (\e.(?&byte)){3} \eb
2074     .sp
2075     The first part of the pattern is a DEFINE group inside which a another group
2076     named "byte" is defined. This matches an individual component of an IPv4
2077     address (a number less than 256). When matching takes place, this part of the
2078 ph10 456 pattern is skipped because DEFINE acts like a false condition. The rest of the
2079     pattern uses references to the named group to match the four dot-separated
2080     components of an IPv4 address, insisting on a word boundary at each end.
2081 nigel 93 .
2082     .SS "Assertion conditions"
2083     .rs
2084     .sp
2085     If the condition is not in any of the above formats, it must be an assertion.
2086 nigel 63 This may be a positive or negative lookahead or lookbehind assertion. Consider
2087     this pattern, again containing non-significant white space, and with the two
2088     alternatives on the second line:
2089 nigel 75 .sp
2090 nigel 63 (?(?=[^a-z]*[a-z])
2091 nigel 75 \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
2092     .sp
2093 nigel 63 The condition is a positive lookahead assertion that matches an optional
2094     sequence of non-letters followed by a letter. In other words, it tests for the
2095     presence of at least one letter in the subject. If a letter is found, the
2096     subject is matched against the first alternative; otherwise it is matched
2097     against the second. This pattern matches strings in one of the two forms
2098     dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
2099 nigel 75 .
2100     .
2101     .\" HTML <a name="comments"></a>
2102 nigel 63 .SH COMMENTS
2103     .rs
2104     .sp
2105 nigel 75 The sequence (?# marks the start of a comment that continues up to the next
2106 nigel 63 closing parenthesis. Nested parentheses are not permitted. The characters
2107     that make up a comment play no part in the pattern matching at all.
2108 nigel 75 .P
2109 nigel 63 If the PCRE_EXTENDED option is set, an unescaped # character outside a
2110 nigel 91 character class introduces a comment that continues to immediately after the
2111     next newline in the pattern.
2112 nigel 75 .
2113     .
2114 nigel 91 .\" HTML <a name="recursion"></a>
2115 nigel 75 .SH "RECURSIVE PATTERNS"
2116 nigel 63 .rs
2117     .sp
2118     Consider the problem of matching a string in parentheses, allowing for
2119     unlimited nested parentheses. Without the use of recursion, the best that can
2120     be done is to use a pattern that matches up to some fixed depth of nesting. It
2121 nigel 93 is not possible to handle an arbitrary nesting depth.
2122     .P
2123     For some time, Perl has provided a facility that allows regular expressions to
2124     recurse (amongst other things). It does this by interpolating Perl code in the
2125     expression at run time, and the code can refer to the expression itself. A Perl
2126     pattern using code interpolation to solve the parentheses problem can be
2127     created like this:
2128 nigel 75 .sp
2129     $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
2130     .sp
2131 nigel 63 The (?p{...}) item interpolates Perl code at run time, and in this case refers
2132 nigel 93 recursively to the pattern in which it appears.
2133 nigel 75 .P
2134 nigel 93 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
2135     supports special syntax for recursion of the entire pattern, and also for
2136     individual subpattern recursion. After its introduction in PCRE and Python,
2137 ph10 453 this kind of recursion was subsequently introduced into Perl at release 5.10.
2138 nigel 75 .P
2139 nigel 93 A special item that consists of (? followed by a number greater than zero and a
2140     closing parenthesis is a recursive call of the subpattern of the given number,
2141 ph10 461 provided that it occurs inside that subpattern. (If not, it is a
2142 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
2143     .\" </a>
2144     "subroutine"
2145     .\"
2146 nigel 93 call, which is described in the next section.) The special item (?R) or (?0) is
2147     a recursive call of the entire regular expression.
2148 nigel 87 .P
2149     This PCRE pattern solves the nested parentheses problem (assume the
2150     PCRE_EXTENDED option is set so that white space is ignored):
2151 nigel 75 .sp
2152 ph10 456 \e( ( [^()]++ | (?R) )* \e)
2153 nigel 75 .sp
2154 nigel 63 First it matches an opening parenthesis. Then it matches any number of
2155     substrings which can either be a sequence of non-parentheses, or a recursive
2156 nigel 87 match of the pattern itself (that is, a correctly parenthesized substring).
2157 ph10 461 Finally there is a closing parenthesis. Note the use of a possessive quantifier
2158 ph10 456 to avoid backtracking into sequences of non-parentheses.
2159 nigel 75 .P
2160 nigel 63 If this were part of a larger pattern, you would not want to recurse the entire
2161     pattern, so instead you could use this:
2162 nigel 75 .sp
2163 ph10 456 ( \e( ( [^()]++ | (?1) )* \e) )
2164 nigel 75 .sp
2165 nigel 63 We have put the pattern into parentheses, and caused the recursion to refer to
2166 ph10 172 them instead of the whole pattern.
2167 ph10 166 .P
2168     In a larger pattern, keeping track of parenthesis numbers can be tricky. This
2169 ph10 456 is made easier by the use of relative references (a Perl 5.10 feature).
2170 ph10 166 Instead of (?1) in the pattern above you can write (?-2) to refer to the second
2171     most recently opened parentheses preceding the recursion. In other words, a
2172     negative number counts capturing parentheses leftwards from the point at which
2173     it is encountered.
2174     .P
2175     It is also possible to refer to subsequently opened parentheses, by writing
2176     references such as (?+2). However, these cannot be recursive because the
2177     reference is not inside the parentheses that are referenced. They are always
2178 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
2179     .\" </a>
2180     "subroutine"
2181     .\"
2182     calls, as described in the next section.
2183 ph10 166 .P
2184     An alternative approach is to use named parentheses instead. The Perl syntax
2185     for this is (?&name); PCRE's earlier syntax (?P>name) is also supported. We
2186     could rewrite the above example as follows:
2187 nigel 75 .sp
2188 ph10 456 (?<pn> \e( ( [^()]++ | (?&pn) )* \e) )
2189 nigel 75 .sp
2190 nigel 93 If there is more than one subpattern with the same name, the earliest one is
2191 ph10 172 used.
2192 ph10 166 .P
2193     This particular example pattern that we have been looking at contains nested
2194 ph10 456 unlimited repeats, and so the use of a possessive quantifier for matching
2195     strings of non-parentheses is important when applying the pattern to strings
2196     that do not match. For example, when this pattern is applied to
2197 nigel 75 .sp
2198 nigel 63 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
2199 nigel 75 .sp
2200 ph10 456 it yields "no match" quickly. However, if a possessive quantifier is not used,
2201 nigel 63 the match runs for a very long time indeed because there are so many different
2202     ways the + and * repeats can carve up the subject, and all have to be tested
2203     before failure can be reported.
2204 nigel 75 .P
2205 ph10 464 At the end of a match, the values of capturing parentheses are those from
2206     the outermost level. If you want to obtain intermediate values, a callout
2207     function can be used (see below and the
2208 nigel 63 .\" HREF
2209 nigel 75 \fBpcrecallout\fP
2210 nigel 63 .\"
2211     documentation). If the pattern above is matched against
2212 nigel 75 .sp
2213 nigel 63 (ab(cd)ef)
2214 nigel 75 .sp
2215 ph10 464 the value for the inner capturing parentheses (numbered 2) is "ef", which is
2216     the last value taken on at the top level. If a capturing subpattern is not
2217     matched at the top level, its final value is unset, even if it is (temporarily)
2218     set at a deeper level.
2219 nigel 75 .P
2220 ph10 464 If there are more than 15 capturing parentheses in a pattern, PCRE has to
2221     obtain extra memory to store data during a recursion, which it does by using
2222     \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no memory can
2223     be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
2224     .P
2225 nigel 63 Do not confuse the (?R) item with the condition (R), which tests for recursion.
2226     Consider this pattern, which matches text in angle brackets, allowing for
2227     arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2228     recursing), whereas any characters are permitted at the outer level.
2229 nigel 75 .sp
2230     < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
2231     .sp
2232 nigel 63 In this pattern, (?(R) is the start of a conditional subpattern, with two
2233     different alternatives for the recursive and non-recursive cases. The (?R) item
2234     is the actual recursive call.
2235 nigel 75 .
2236     .
2237 ph10 453 .\" HTML <a name="recursiondifference"></a>
2238     .SS "Recursion difference from Perl"
2239     .rs
2240     .sp
2241     In PCRE (like Python, but unlike Perl), a recursive subpattern call is always
2242     treated as an atomic group. That is, once it has matched some of the subject
2243     string, it is never re-entered, even if it contains untried alternatives and
2244 ph10 461 there is a subsequent matching failure. This can be illustrated by the
2245     following pattern, which purports to match a palindromic string that contains
2246 ph10 453 an odd number of characters (for example, "a", "aba", "abcba", "abcdcba"):
2247     .sp
2248     ^(.|(.)(?1)\e2)$
2249     .sp
2250 ph10 461 The idea is that it either matches a single character, or two identical
2251     characters surrounding a sub-palindrome. In Perl, this pattern works; in PCRE
2252 ph10 453 it does not if the pattern is longer than three characters. Consider the
2253     subject string "abcba":
2254     .P
2255 ph10 461 At the top level, the first character is matched, but as it is not at the end
2256 ph10 453 of the string, the first alternative fails; the second alternative is taken
2257     and the recursion kicks in. The recursive call to subpattern 1 successfully
2258     matches the next character ("b"). (Note that the beginning and end of line
2259     tests are not part of the recursion).
2260     .P
2261     Back at the top level, the next character ("c") is compared with what
2262 ph10 461 subpattern 2 matched, which was "a". This fails. Because the recursion is
2263 ph10 453 treated as an atomic group, there are now no backtracking points, and so the
2264     entire match fails. (Perl is able, at this point, to re-enter the recursion and
2265     try the second alternative.) However, if the pattern is written with the
2266     alternatives in the other order, things are different:
2267     .sp
2268     ^((.)(?1)\e2|.)$
2269     .sp
2270 ph10 461 This time, the recursing alternative is tried first, and continues to recurse
2271     until it runs out of characters, at which point the recursion fails. But this
2272     time we do have another alternative to try at the higher level. That is the big
2273 ph10 453 difference: in the previous case the remaining alternative is at a deeper
2274     recursion level, which PCRE cannot use.
2275     .P
2276 ph10 461 To change the pattern so that matches all palindromic strings, not just those
2277 ph10 453 with an odd number of characters, it is tempting to change the pattern to this:
2278     .sp
2279     ^((.)(?1)\e2|.?)$
2280     .sp
2281 ph10 461 Again, this works in Perl, but not in PCRE, and for the same reason. When a
2282     deeper recursion has matched a single character, it cannot be entered again in
2283     order to match an empty string. The solution is to separate the two cases, and
2284 ph10 453 write out the odd and even cases as alternatives at the higher level:
2285     .sp
2286     ^(?:((.)(?1)\e2|)|((.)(?3)\e4|.))
2287 ph10 461 .sp
2288     If you want to match typical palindromic phrases, the pattern has to ignore all
2289 ph10 453 non-word characters, which can be done like this:
2290     .sp
2291 ph10 461 ^\eW*+(?:((.)\eW*+(?1)\eW*+\e2|)|((.)\eW*+(?3)\eW*+\e4|\eW*+.\eW*+))\eW*+$
2292 ph10 453 .sp
2293 ph10 461 If run with the PCRE_CASELESS option, this pattern matches phrases such as "A
2294     man, a plan, a canal: Panama!" and it works well in both PCRE and Perl. Note
2295     the use of the possessive quantifier *+ to avoid backtracking into sequences of
2296 ph10 453 non-word characters. Without this, PCRE takes a great deal longer (ten times or
2297     more) to match typical phrases, and Perl takes so long that you think it has
2298     gone into a loop.
2299 ph10 456 .P
2300     \fBWARNING\fP: The palindrome-matching patterns above work only if the subject
2301     string does not start with a palindrome that is shorter than the entire string.
2302     For example, although "abcba" is correctly matched, if the subject is "ababa",
2303     PCRE finds the palindrome "aba" at the start, then fails at top level because
2304     the end of the string does not follow. Once again, it cannot jump back into the
2305     recursion to try other alternatives, so the entire match fails.
2306 ph10 453 .
2307     .
2308 nigel 63 .\" HTML <a name="subpatternsassubroutines"></a>
2309 nigel 75 .SH "SUBPATTERNS AS SUBROUTINES"
2310 nigel 63 .rs
2311     .sp
2312     If the syntax for a recursive subpattern reference (either by number or by
2313     name) is used outside the parentheses to which it refers, it operates like a
2314 nigel 93 subroutine in a programming language. The "called" subpattern may be defined
2315 ph10 166 before or after the reference. A numbered reference can be absolute or
2316     relative, as in these examples:
2317 nigel 75 .sp
2318 ph10 166 (...(absolute)...)...(?2)...
2319     (...(relative)...)...(?-1)...
2320 ph10 172 (...(?+1)...(relative)...
2321 ph10 166 .sp
2322     An earlier example pointed out that the pattern
2323     .sp
2324 nigel 75 (sens|respons)e and \e1ibility
2325     .sp
2326 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
2327     "sense and responsibility". If instead the pattern
2328 nigel 75 .sp
2329 nigel 63 (sens|respons)e and (?1)ibility
2330 nigel 75 .sp
2331 nigel 63 is used, it does match "sense and responsibility" as well as the other two
2332 nigel 93 strings. Another example is given in the discussion of DEFINE above.
2333 nigel 87 .P
2334 ph10 464 Like recursive subpatterns, a subroutine call is always treated as an atomic
2335 nigel 87 group. That is, once it has matched some of the subject string, it is never
2336     re-entered, even if it contains untried alternatives and there is a subsequent
2337 ph10 469 matching failure. Any capturing parentheses that are set during the subroutine
2338 ph10 464 call revert to their previous values afterwards.
2339 nigel 93 .P
2340     When a subpattern is used as a subroutine, processing options such as
2341     case-independence are fixed when the subpattern is defined. They cannot be
2342     changed for different calls. For example, consider this pattern:
2343     .sp
2344 ph10 166 (abc)(?i:(?-1))
2345 nigel 93 .sp
2346     It matches "abcabc". It does not match "abcABC" because the change of
2347     processing option does not affect the called subpattern.
2348 nigel 75 .
2349     .
2350 ph10 333 .\" HTML <a name="onigurumasubroutines"></a>
2351     .SH "ONIGURUMA SUBROUTINE SYNTAX"
2352     .rs
2353     .sp
2354 ph10 345 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
2355     a number enclosed either in angle brackets or single quotes, is an alternative
2356     syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2357 ph10 333 are two of the examples used above, rewritten using this syntax:
2358     .sp
2359     (?<pn> \e( ( (?>[^()]+) | \eg<pn> )* \e) )
2360     (sens|respons)e and \eg'1'ibility
2361     .sp
2362 ph10 345 PCRE supports an extension to Oniguruma: if a number is preceded by a
2363 ph10 333 plus or a minus sign it is taken as a relative reference. For example:
2364     .sp
2365     (abc)(?i:\eg<-1>)
2366     .sp
2367 ph10 345 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
2368 ph10 333 synonymous. The former is a back reference; the latter is a subroutine call.
2369     .
2370     .
2371 nigel 63 .SH CALLOUTS
2372     .rs
2373     .sp
2374     Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2375     code to be obeyed in the middle of matching a regular expression. This makes it
2376     possible, amongst other things, to extract different substrings that match the
2377     same pair of parentheses when there is a repetition.
2378 nigel 75 .P
2379 nigel 63 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2380     code. The feature is called "callout". The caller of PCRE provides an external
2381 nigel 75 function by putting its entry point in the global variable \fIpcre_callout\fP.
2382 nigel 63 By default, this variable contains NULL, which disables all calling out.
2383 nigel 75 .P
2384 nigel 63 Within a regular expression, (?C) indicates the points at which the external
2385     function is to be called. If you want to identify different callout points, you
2386     can put a number less than 256 after the letter C. The default value is zero.
2387     For example, this pattern has two callout points:
2388 nigel 75 .sp
2389 ph10 155 (?C1)abc(?C2)def
2390 nigel 75 .sp
2391     If the PCRE_AUTO_CALLOUT flag is passed to \fBpcre_compile()\fP, callouts are
2392     automatically installed before each item in the pattern. They are all numbered
2393     255.
2394     .P
2395     During matching, when PCRE reaches a callout point (and \fIpcre_callout\fP is
2396 nigel 63 set), the external function is called. It is provided with the number of the
2397 nigel 75 callout, the position in the pattern, and, optionally, one item of data
2398     originally supplied by the caller of \fBpcre_exec()\fP. The callout function
2399     may cause matching to proceed, to backtrack, or to fail altogether. A complete
2400     description of the interface to the callout function is given in the
2401 nigel 63 .\" HREF
2402 nigel 75 \fBpcrecallout\fP
2403 nigel 63 .\"
2404     documentation.
2405 nigel 93 .
2406     .
2407 ph10 510 .\" HTML <a name="backtrackcontrol"></a>
2408 ph10 235 .SH "BACKTRACKING CONTROL"
2409 ph10 210 .rs
2410     .sp
2411 ph10 211 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2412 ph10 210 are described in the Perl documentation as "experimental and subject to change
2413 ph10 211 or removal in a future version of Perl". It goes on to say: "Their usage in
2414     production code should be noted to avoid problems during upgrades." The same
2415 ph10 210 remarks apply to the PCRE features described in this section.
2416     .P
2417 ph10 341 Since these verbs are specifically related to backtracking, most of them can be
2418     used only when the pattern is to be matched using \fBpcre_exec()\fP, which uses
2419 ph10 345 a backtracking algorithm. With the exception of (*FAIL), which behaves like a
2420 ph10 341 failing negative assertion, they cause an error if encountered by
2421 ph10 210 \fBpcre_dfa_exec()\fP.
2422     .P
2423 ph10 469 If any of these verbs are used in an assertion or subroutine subpattern
2424 ph10 464 (including recursive subpatterns), their effect is confined to that subpattern;
2425     it does not extend to the surrounding pattern. Note that such subpatterns are
2426     processed as anchored at the point where they are tested.
2427 ph10 445 .P
2428 ph10 211 The new verbs make use of what was previously invalid syntax: an opening
2429 ph10 510 parenthesis followed by an asterisk. They are generally of the form
2430 ph10 512 (*VERB) or (*VERB:NAME). Some may take either form, with differing behaviour,
2431 ph10 510 depending on whether or not an argument is present. An name is a sequence of
2432     letters, digits, and underscores. If the name is empty, that is, if the closing
2433     parenthesis immediately follows the colon, the effect is as if the colon were
2434     not there. Any number of these verbs may occur in a pattern.
2435     .P
2436 ph10 512 PCRE contains some optimizations that are used to speed up matching by running
2437     some checks at the start of each match attempt. For example, it may know the
2438     minimum length of matching subject, or that a particular character must be
2439     present. When one of these optimizations suppresses the running of a match, any
2440     included backtracking verbs will not, of course, be processed. You can suppress
2441     the start-of-match optimizations by setting the PCRE_NO_START_OPTIMIZE option
2442 ph10 510 when calling \fBpcre_exec()\fP.
2443 ph10 210 .
2444 ph10 510 .
2445 ph10 210 .SS "Verbs that act immediately"
2446     .rs
2447     .sp
2448 ph10 512 The following verbs act as soon as they are encountered. They may not be
2449 ph10 510 followed by a name.
2450 ph10 210 .sp
2451     (*ACCEPT)
2452     .sp
2453     This verb causes the match to end successfully, skipping the remainder of the
2454 ph10 211 pattern. When inside a recursion, only the innermost pattern is ended
2455 ph10 456 immediately. If (*ACCEPT) is inside capturing parentheses, the data so far is
2456     captured. (This feature was added to PCRE at release 8.00.) For example:
2457 ph10 210 .sp
2458 ph10 447 A((?:A|B(*ACCEPT)|C)D)
2459 ph10 210 .sp
2460 ph10 461 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is captured by
2461 ph10 447 the outer parentheses.
2462 ph10 210 .sp
2463     (*FAIL) or (*F)
2464     .sp
2465 ph10 211 This verb causes the match to fail, forcing backtracking to occur. It is
2466 ph10 210 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2467     probably useful only when combined with (?{}) or (??{}). Those are, of course,
2468     Perl features that are not present in PCRE. The nearest equivalent is the
2469     callout feature, as for example in this pattern:
2470     .sp
2471     a+(?C)(*FAIL)
2472     .sp
2473 ph10 211 A match with the string "aaaa" always fails, but the callout is taken before
2474     each backtrack happens (in this example, 10 times).
2475 ph10 210 .
2476 ph10 510 .
2477     .SS "Recording which path was taken"
2478     .rs
2479     .sp
2480 ph10 512 There is one verb whose main purpose is to track how a match was arrived at,
2481     though it also has a secondary use in conjunction with advancing the match
2482 ph10 510 starting point (see (*SKIP) below).
2483     .sp
2484     (*MARK:NAME) or (*:NAME)
2485     .sp
2486     A name is always required with this verb. There may be as many instances of
2487     (*MARK) as you like in a pattern, and their names do not have to be unique.
2488     .P
2489     When a match succeeds, the name of the last-encountered (*MARK) is passed back
2490     to the caller via the \fIpcre_extra\fP data structure, as described in the
2491     .\" HTML <a href="pcreapi.html#extradata">
2492     .\" </a>
2493     section on \fIpcre_extra\fP
2494     .\"
2495 ph10 512 in the
2496 ph10 510 .\" HREF
2497     \fBpcreapi\fP
2498     .\"
2499     documentation. No data is returned for a partial match. Here is an example of
2500     \fBpcretest\fP output, where the /K modifier requests the retrieval and
2501     outputting of (*MARK) data:
2502     .sp
2503     /X(*MARK:A)Y|X(*MARK:B)Z/K
2504     XY
2505     0: XY
2506     MK: A
2507     XZ
2508     0: XZ
2509     MK: B
2510     .sp
2511 ph10 512 The (*MARK) name is tagged with "MK:" in this output, and in this example it
2512     indicates which of the two alternatives matched. This is a more efficient way
2513 ph10 510 of obtaining this information than putting each alternative in its own
2514     capturing parentheses.
2515     .P
2516     A name may also be returned after a failed match if the final path through the
2517     pattern involves (*MARK). However, unless (*MARK) used in conjunction with
2518     (*COMMIT), this is unlikely to happen for an unanchored pattern because, as the
2519     starting point for matching is advanced, the final check is often with an empty
2520     string, causing a failure before (*MARK) is reached. For example:
2521     .sp
2522     /X(*MARK:A)Y|X(*MARK:B)Z/K
2523     XP
2524     No match
2525     .sp
2526     There are three potential starting points for this match (starting with X,
2527 ph10 512 starting with P, and with an empty string). If the pattern is anchored, the
2528 ph10 510 result is different:
2529     .sp
2530     /^X(*MARK:A)Y|^X(*MARK:B)Z/K
2531     XP
2532     No match, mark = B
2533     .sp
2534 ph10 512 PCRE's start-of-match optimizations can also interfere with this. For example,
2535     if, as a result of a call to \fBpcre_study()\fP, it knows the minimum
2536 ph10 510 subject length for a match, a shorter subject will not be scanned at all.
2537     .P
2538 ph10 512 Note that similar anomalies (though different in detail) exist in Perl, no
2539     doubt for the same reasons. The use of (*MARK) data after a failed match of an
2540 ph10 510 unanchored pattern is not recommended, unless (*COMMIT) is involved.
2541     .
2542     .
2543 ph10 210 .SS "Verbs that act after backtracking"
2544     .rs
2545     .sp
2546 ph10 211 The following verbs do nothing when they are encountered. Matching continues
2547 ph10 510 with what follows, but if there is no subsequent match, causing a backtrack to
2548     the verb, a failure is forced. That is, backtracking cannot pass to the left of
2549     the verb. However, when one of these verbs appears inside an atomic group, its
2550     effect is confined to that group, because once the group has been matched,
2551     there is never any backtracking into it. In this situation, backtracking can
2552 ph10 512 "jump back" to the left of the entire atomic group. (Remember also, as stated
2553 ph10 510 above, that this localization also applies in subroutine calls and assertions.)
2554     .P
2555     These verbs differ in exactly what kind of failure occurs when backtracking
2556     reaches them.
2557 ph10 210 .sp
2558     (*COMMIT)
2559     .sp
2560 ph10 510 This verb, which may not be followed by a name, causes the whole match to fail
2561     outright if the rest of the pattern does not match. Even if the pattern is
2562     unanchored, no further attempts to find a match by advancing the starting point
2563     take place. Once (*COMMIT) has been passed, \fBpcre_exec()\fP is committed to
2564     finding a match at the current starting point, or not at all. For example:
2565 ph10 210 .sp
2566     a+(*COMMIT)b
2567     .sp
2568 ph10 211 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2569 ph10 512 dynamic anchor, or "I've started, so I must finish." The name of the most
2570     recently passed (*MARK) in the path is passed back when (*COMMIT) forces a
2571 ph10 510 match failure.
2572     .P
2573 ph10 512 Note that (*COMMIT) at the start of a pattern is not the same as an anchor,
2574     unless PCRE's start-of-match optimizations are turned off, as shown in this
2575 ph10 510 \fBpcretest\fP example:
2576 ph10 210 .sp
2577 ph10 510 /(*COMMIT)abc/
2578     xyzabc
2579     0: abc
2580     xyzabc\eY
2581     No match
2582 ph10 210 .sp
2583 ph10 512 PCRE knows that any match must start with "a", so the optimization skips along
2584 ph10 510 the subject to "a" before running the first match attempt, which succeeds. When
2585     the optimization is disabled by the \eY escape in the second subject, the match
2586 ph10 512 starts at "x" and so the (*COMMIT) causes it to fail without trying any other
2587 ph10 510 starting points.
2588 ph10 210 .sp
2589 ph10 510 (*PRUNE) or (*PRUNE:NAME)
2590     .sp
2591 ph10 512 This verb causes the match to fail at the current starting position in the
2592 ph10 510 subject if the rest of the pattern does not match. If the pattern is
2593     unanchored, the normal "bumpalong" advance to the next starting character then
2594     happens. Backtracking can occur as usual to the left of (*PRUNE), before it is
2595     reached, or when matching to the right of (*PRUNE), but if there is no match to
2596     the right, backtracking cannot cross (*PRUNE). In simple cases, the use of
2597     (*PRUNE) is just an alternative to an atomic group or possessive quantifier,
2598     but there are some uses of (*PRUNE) that cannot be expressed in any other way.
2599     The behaviour of (*PRUNE:NAME) is the same as (*MARK:NAME)(*PRUNE) when the
2600     match fails completely; the name is passed back if this is the final attempt.
2601     (*PRUNE:NAME) does not pass back a name if the match succeeds. In an anchored
2602     pattern (*PRUNE) has the same effect as (*COMMIT).
2603     .sp
2604 ph10 210 (*SKIP)
2605     .sp
2606 ph10 510 This verb, when given without a name, is like (*PRUNE), except that if the
2607     pattern is unanchored, the "bumpalong" advance is not to the next character,
2608     but to the position in the subject where (*SKIP) was encountered. (*SKIP)
2609     signifies that whatever text was matched leading up to it cannot be part of a
2610     successful match. Consider:
2611 ph10 210 .sp
2612     a+(*SKIP)b
2613     .sp
2614 ph10 211 If the subject is "aaaac...", after the first match attempt fails (starting at
2615 ph10 210 the first character in the string), the starting point skips on to start the
2616 ph10 211 next attempt at "c". Note that a possessive quantifer does not have the same
2617 ph10 456 effect as this example; although it would suppress backtracking during the
2618 ph10 210 first match attempt, the second attempt would start at the second character
2619     instead of skipping on to "c".
2620     .sp
2621 ph10 510 (*SKIP:NAME)
2622 ph10 211 .sp
2623 ph10 512 When (*SKIP) has an associated name, its behaviour is modified. If the
2624     following pattern fails to match, the previous path through the pattern is
2625     searched for the most recent (*MARK) that has the same name. If one is found,
2626     the "bumpalong" advance is to the subject position that corresponds to that
2627     (*MARK) instead of to where (*SKIP) was encountered. If no (*MARK) with a
2628     matching name is found, normal "bumpalong" of one character happens (the
2629 ph10 510 (*SKIP) is ignored).
2630     .sp
2631     (*THEN) or (*THEN:NAME)
2632     .sp
2633 ph10 210 This verb causes a skip to the next alternation if the rest of the pattern does
2634     not match. That is, it cancels pending backtracking, but only within the
2635     current alternation. Its name comes from the observation that it can be used
2636     for a pattern-based if-then-else block:
2637     .sp
2638     ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
2639     .sp
2640 ph10 211 If the COND1 pattern matches, FOO is tried (and possibly further items after
2641 ph10 210 the end of the group if FOO succeeds); on failure the matcher skips to the
2642 ph10 510 second alternative and tries COND2, without backtracking into COND1. The
2643     behaviour of (*THEN:NAME) is exactly the same as (*MARK:NAME)(*THEN) if the
2644     overall match fails. If (*THEN) is not directly inside an alternation, it acts
2645     like (*PRUNE).
2646 ph10 210 .
2647     .
2648 nigel 93 .SH "SEE ALSO"
2649     .rs
2650     .sp
2651 ph10 461 \fBpcreapi\fP(3), \fBpcrecallout\fP(3), \fBpcrematching\fP(3),
2652 ph10 456 \fBpcresyntax\fP(3), \fBpcre\fP(3).
2653 ph10 99 .
2654     .
2655     .SH AUTHOR
2656     .rs
2657     .sp
2658     .nf
2659     Philip Hazel
2660     University Computing Service
2661     Cambridge CB2 3QH, England.
2662     .fi
2663     .
2664     .
2665     .SH REVISION
2666     .rs
2667     .sp
2668     .nf
2669 ph10 518 Last updated: 18 May 2010
2670 ph10 488 Copyright (c) 1997-2010 University of Cambridge.
2671 ph10 99 .fi

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12