ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 745 - (hide annotations) (download)
Mon Nov 14 11:41:03 2011 UTC (3 years, 5 months ago) by ph10
File size: 120327 byte(s)
Small tidies, and documentation update for JavaScript \x, \u, \U support.

1 nigel 79 .TH PCREPATTERN 3
2 nigel 63 .SH NAME
3     PCRE - Perl-compatible regular expressions
5 nigel 63 .rs
6     .sp
7 ph10 208 The syntax and semantics of the regular expressions that are supported by PCRE
8     are described in detail below. There is a quick-reference syntax summary in the
9     .\" HREF
10     \fBpcresyntax\fP
11     .\"
12 ph10 333 page. PCRE tries to match Perl syntax and semantics as closely as it can. PCRE
13     also supports some alternative regular expression syntax (which does not
14     conflict with the Perl syntax) in order to provide some compatibility with
15     regular expressions in Python, .NET, and Oniguruma.
16     .P
17     Perl's regular expressions are described in its own documentation, and
18 ph10 208 regular expressions in general are covered in a number of books, some of which
19     have copious examples. Jeffrey Friedl's "Mastering Regular Expressions",
20     published by O'Reilly, covers regular expressions in great detail. This
21     description of PCRE's regular expressions is intended as reference material.
22 nigel 75 .P
23     The original operation of PCRE was on strings of one-byte characters. However,
24 ph10 461 there is now also support for UTF-8 character strings. To use this,
25 ph10 456 PCRE must be built to include UTF-8 support, and you must call
26     \fBpcre_compile()\fP or \fBpcre_compile2()\fP with the PCRE_UTF8 option. There
27     is also a special sequence that can be given at the start of a pattern:
28 ph10 412 .sp
29     (*UTF8)
30 ph10 416 .sp
31 ph10 412 Starting a pattern with this sequence is equivalent to setting the PCRE_UTF8
32     option. This feature is not Perl-compatible. How setting UTF-8 mode affects
33     pattern matching is mentioned in several places below. There is also a summary
34     of UTF-8 features in the
35 nigel 63 .\" HREF
36 ph10 678 \fBpcreunicode\fP
37 nigel 63 .\"
38     page.
39 nigel 75 .P
40 ph10 535 Another special sequence that may appear at the start of a pattern or in
41 ph10 518 combination with (*UTF8) is:
42     .sp
43     (*UCP)
44     .sp
45 ph10 535 This has the same effect as setting the PCRE_UCP option: it causes sequences
46     such as \ed and \ew to use Unicode properties to determine character types,
47     instead of recognizing only characters with codes less than 128 via a lookup
48 ph10 518 table.
49     .P
50 ph10 576 If a pattern starts with (*NO_START_OPT), it has the same effect as setting the
51 ph10 579 PCRE_NO_START_OPTIMIZE option either at compile or matching time. There are
52     also some more of these special sequences that are concerned with the handling
53 ph10 576 of newlines; they are described below.
54     .P
55 nigel 77 The remainder of this document discusses the patterns that are supported by
56     PCRE when its main matching function, \fBpcre_exec()\fP, is used.
57     From release 6.0, PCRE offers a second matching function,
58     \fBpcre_dfa_exec()\fP, which matches using a different algorithm that is not
59 ph10 172 Perl-compatible. Some of the features discussed below are not available when
60 ph10 168 \fBpcre_dfa_exec()\fP is used. The advantages and disadvantages of the
61     alternative function, and how it differs from the normal function, are
62     discussed in the
63 nigel 77 .\" HREF
64     \fBpcrematching\fP
65     .\"
66     page.
67 nigel 93 .
68     .
69 ph10 556 .\" HTML <a name="newlines"></a>
71     .rs
72     .sp
73     PCRE supports five different conventions for indicating line breaks in
74     strings: a single CR (carriage return) character, a single LF (linefeed)
75     character, the two-character sequence CRLF, any of the three preceding, or any
76     Unicode newline sequence. The
77     .\" HREF
78     \fBpcreapi\fP
79     .\"
80     page has
81     .\" HTML <a href="pcreapi.html#newlines">
82     .\" </a>
83     further discussion
84     .\"
85     about newlines, and shows how to set the newline convention in the
86     \fIoptions\fP arguments for the compiling and matching functions.
87     .P
88     It is also possible to specify a newline convention by starting a pattern
89     string with one of the following five sequences:
90     .sp
91     (*CR) carriage return
92     (*LF) linefeed
93     (*CRLF) carriage return, followed by linefeed
94     (*ANYCRLF) any of the three above
95     (*ANY) all Unicode newline sequences
96     .sp
97 ph10 461 These override the default and the options given to \fBpcre_compile()\fP or
98 ph10 456 \fBpcre_compile2()\fP. For example, on a Unix system where LF is the default
99     newline sequence, the pattern
100 ph10 227 .sp
101     (*CR)a.b
102     .sp
103     changes the convention to CR. That pattern matches "a\enb" because LF is no
104     longer a newline. Note that these special settings, which are not
105     Perl-compatible, are recognized only at the very start of a pattern, and that
106 ph10 231 they must be in upper case. If more than one of them is present, the last one
107     is used.
108     .P
109 ph10 514 The newline convention affects the interpretation of the dot metacharacter when
110     PCRE_DOTALL is not set, and also the behaviour of \eN. However, it does not
111     affect what the \eR escape sequence matches. By default, this is any Unicode
112     newline sequence, for Perl compatibility. However, this can be changed; see the
113     description of \eR in the section entitled
114 ph10 231 .\" HTML <a href="#newlineseq">
115     .\" </a>
116     "Newline sequences"
117     .\"
118 ph10 247 below. A change of \eR setting can be combined with a change of newline
119 ph10 246 convention.
120 ph10 227 .
121     .
123     .rs
124     .sp
125 nigel 63 A regular expression is a pattern that is matched against a subject string from
126     left to right. Most characters stand for themselves in a pattern, and match the
127     corresponding characters in the subject. As a trivial example, the pattern
128 nigel 75 .sp
129 nigel 63 The quick brown fox
130 nigel 75 .sp
131 nigel 77 matches a portion of a subject string that is identical to itself. When
132     caseless matching is specified (the PCRE_CASELESS option), letters are matched
133     independently of case. In UTF-8 mode, PCRE always understands the concept of
134     case for characters whose values are less than 128, so caseless matching is
135     always possible. For characters with higher values, the concept of case is
136     supported if PCRE is compiled with Unicode property support, but not otherwise.
137     If you want to use caseless matching for characters 128 and above, you must
138     ensure that PCRE is compiled with Unicode property support as well as with
139     UTF-8 support.
140     .P
141     The power of regular expressions comes from the ability to include alternatives
142     and repetitions in the pattern. These are encoded in the pattern by the use of
143 nigel 75 \fImetacharacters\fP, which do not stand for themselves but instead are
144 nigel 63 interpreted in some special way.
145 nigel 75 .P
146     There are two different sets of metacharacters: those that are recognized
147 nigel 63 anywhere in the pattern except within square brackets, and those that are
148 nigel 93 recognized within square brackets. Outside square brackets, the metacharacters
149     are as follows:
150 nigel 75 .sp
151     \e general escape character with several uses
152 nigel 63 ^ assert start of string (or line, in multiline mode)
153     $ assert end of string (or line, in multiline mode)
154     . match any character except newline (by default)
155     [ start character class definition
156     | start of alternative branch
157     ( start subpattern
158     ) end subpattern
159     ? extends the meaning of (
160     also 0 or 1 quantifier
161     also quantifier minimizer
162     * 0 or more quantifier
163     + 1 or more quantifier
164     also "possessive quantifier"
165     { start min/max quantifier
166 nigel 75 .sp
167 nigel 63 Part of a pattern that is in square brackets is called a "character class". In
168 nigel 75 a character class the only metacharacters are:
169     .sp
170     \e general escape character
171 nigel 63 ^ negate the class, but only if the first character
172     - indicates character range
173 nigel 75 .\" JOIN
174 nigel 63 [ POSIX character class (only if followed by POSIX
175     syntax)
176     ] terminates the character class
177 nigel 75 .sp
178     The following sections describe the use of each of the metacharacters.
179     .
180 nigel 93 .
181 nigel 63 .SH BACKSLASH
182     .rs
183     .sp
184     The backslash character has several uses. Firstly, if it is followed by a
185 ph10 574 character that is not a number or a letter, it takes away any special meaning
186     that character may have. This use of backslash as an escape character applies
187 ph10 579 both inside and outside character classes.
188 nigel 75 .P
189     For example, if you want to match a * character, you write \e* in the pattern.
190 nigel 63 This escaping action applies whether or not the following character would
191 nigel 75 otherwise be interpreted as a metacharacter, so it is always safe to precede a
192     non-alphanumeric with backslash to specify that it stands for itself. In
193     particular, if you want to match a backslash, you write \e\e.
194     .P
195 ph10 574 In UTF-8 mode, only ASCII numbers and letters have any special meaning after a
196 ph10 579 backslash. All other characters (in particular, those whose codepoints are
197 ph10 574 greater than 127) are treated as literals.
198     .P
199 nigel 63 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
200     pattern (other than in a character class) and characters between a # outside
201 nigel 91 a character class and the next newline are ignored. An escaping backslash can
202     be used to include a whitespace or # character as part of the pattern.
203 nigel 75 .P
204 nigel 63 If you want to remove the special meaning from a sequence of characters, you
205 nigel 75 can do so by putting them between \eQ and \eE. This is different from Perl in
206     that $ and @ are handled as literals in \eQ...\eE sequences in PCRE, whereas in
207 nigel 63 Perl, $ and @ cause variable interpolation. Note the following examples:
208 nigel 75 .sp
209 nigel 63 Pattern PCRE matches Perl matches
210 nigel 75 .sp
211     .\" JOIN
212     \eQabc$xyz\eE abc$xyz abc followed by the
213 nigel 63 contents of $xyz
214 nigel 75 \eQabc\e$xyz\eE abc\e$xyz abc\e$xyz
215     \eQabc\eE\e$\eQxyz\eE abc$xyz abc$xyz
216     .sp
217     The \eQ...\eE sequence is recognized both inside and outside character classes.
218 ph10 654 An isolated \eE that is not preceded by \eQ is ignored. If \eQ is not followed
219     by \eE later in the pattern, the literal interpretation continues to the end of
220 ph10 607 the pattern (that is, \eE is assumed at the end). If the isolated \eQ is inside
221     a character class, this causes an error, because the character class is not
222     terminated.
223 nigel 75 .
224     .
225     .\" HTML <a name="digitsafterbackslash"></a>
226     .SS "Non-printing characters"
227     .rs
228     .sp
229 nigel 63 A second use of backslash provides a way of encoding non-printing characters
230     in patterns in a visible manner. There is no restriction on the appearance of
231     non-printing characters, apart from the binary zero that terminates a pattern,
232 ph10 456 but when a pattern is being prepared by text editing, it is often easier to use
233     one of the following escape sequences than the binary character it represents:
234 nigel 75 .sp
235     \ea alarm, that is, the BEL character (hex 07)
236 ph10 574 \ecx "control-x", where x is any ASCII character
237 nigel 75 \ee escape (hex 1B)
238     \ef formfeed (hex 0C)
239 ph10 227 \en linefeed (hex 0A)
240 nigel 75 \er carriage return (hex 0D)
241     \et tab (hex 09)
242 ph10 488 \eddd character with octal code ddd, or back reference
243 nigel 75 \exhh character with hex code hh
244 ph10 745 \ex{hhh..} character with hex code hhh.. (non-JavaScript mode)
245     \euhhhh character with hex code hhhh (JavaScript mode only)
246 nigel 75 .sp
247     The precise effect of \ecx is as follows: if x is a lower case letter, it
248 nigel 63 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
249 ph10 574 Thus \ecz becomes hex 1A (z is 7A), but \ec{ becomes hex 3B ({ is 7B), while
250 ph10 579 \ec; becomes hex 7B (; is 3B). If the byte following \ec has a value greater
251     than 127, a compile-time error occurs. This locks out non-ASCII characters in
252     both byte mode and UTF-8 mode. (When PCRE is compiled in EBCDIC mode, all byte
253     values are valid. A lower case letter is converted to upper case, and then the
254 ph10 574 0xc0 bits are flipped.)
255 nigel 75 .P
256 ph10 745 By default, after \ex, from zero to two hexadecimal digits are read (letters
257     can be in upper or lower case). Any number of hexadecimal digits may appear
258     between \ex{ and }, but the value of the character code must be less than 256
259     in non-UTF-8 mode, and less than 2**31 in UTF-8 mode. That is, the maximum
260     value in hexadecimal is 7FFFFFFF. Note that this is bigger than the largest
261     Unicode code point, which is 10FFFF.
262 nigel 75 .P
263 ph10 211 If characters other than hexadecimal digits appear between \ex{ and }, or if
264     there is no terminating }, this form of escape is not recognized. Instead, the
265     initial \ex will be interpreted as a basic hexadecimal escape, with no
266     following digits, giving a character whose value is zero.
267     .P
268 ph10 745 If the PCRE_JAVASCRIPT_COMPAT option is set, the interpretation of \ex is
269     as just described only when it is followed by two hexadecimal digits.
270     Otherwise, it matches a literal "x" character. In JavaScript mode, support for
271     code points greater than 256 is provided by \eu, which must be followed by
272     four hexadecimal digits; otherwise it matches a literal "u" character.
273     .P
274 nigel 63 Characters whose value is less than 256 can be defined by either of the two
275 ph10 745 syntaxes for \ex (or by \eu in JavaScript mode). There is no difference in the
276     way they are handled. For example, \exdc is exactly the same as \ex{dc} (or
277     \eu00dc in JavaScript mode).
278 nigel 75 .P
279 nigel 91 After \e0 up to two further octal digits are read. If there are fewer than two
280     digits, just those that are present are used. Thus the sequence \e0\ex\e07
281     specifies two binary zeros followed by a BEL character (code value 7). Make
282     sure you supply two digits after the initial zero if the pattern character that
283     follows is itself an octal digit.
284 nigel 75 .P
285 nigel 63 The handling of a backslash followed by a digit other than 0 is complicated.
286     Outside a character class, PCRE reads it and any following digits as a decimal
287     number. If the number is less than 10, or if there have been at least that many
288     previous capturing left parentheses in the expression, the entire sequence is
289 nigel 75 taken as a \fIback reference\fP. A description of how this works is given
290     .\" HTML <a href="#backreferences">
291     .\" </a>
292     later,
293     .\"
294     following the discussion of
295     .\" HTML <a href="#subpattern">
296     .\" </a>
297     parenthesized subpatterns.
298     .\"
299     .P
300 nigel 63 Inside a character class, or if the decimal number is greater than 9 and there
301     have not been that many capturing subpatterns, PCRE re-reads up to three octal
302 nigel 93 digits following the backslash, and uses them to generate a data character. Any
303 nigel 91 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
304     character specified in octal must be less than \e400. In UTF-8 mode, values up
305     to \e777 are permitted. For example:
306 nigel 75 .sp
307     \e040 is another way of writing a space
308     .\" JOIN
309     \e40 is the same, provided there are fewer than 40
310 nigel 63 previous capturing subpatterns
311 nigel 75 \e7 is always a back reference
312     .\" JOIN
313     \e11 might be a back reference, or another way of
314 nigel 63 writing a tab
315 nigel 75 \e011 is always a tab
316     \e0113 is a tab followed by the character "3"
317     .\" JOIN
318     \e113 might be a back reference, otherwise the
319 nigel 63 character with octal code 113
320 nigel 75 .\" JOIN
321     \e377 might be a back reference, otherwise
322 nigel 63 the byte consisting entirely of 1 bits
323 nigel 75 .\" JOIN
324     \e81 is either a back reference, or a binary zero
325 nigel 63 followed by the two characters "8" and "1"
326 nigel 75 .sp
327 nigel 63 Note that octal values of 100 or greater must not be introduced by a leading
328     zero, because no more than three octal digits are ever read.
329 nigel 75 .P
330 nigel 91 All the sequences that define a single character value can be used both inside
331     and outside character classes. In addition, inside a character class, the
332 ph10 513 sequence \eb is interpreted as the backspace character (hex 08). The sequences
333 ph10 514 \eB, \eN, \eR, and \eX are not special inside a character class. Like any other
334 ph10 513 unrecognized escape sequences, they are treated as the literal characters "B",
335 ph10 514 "N", "R", and "X" by default, but cause an error if the PCRE_EXTRA option is
336     set. Outside a character class, these sequences have different meanings.
337 nigel 75 .
338     .
339 ph10 745 .SS "Unsupported escape sequences"
340     .rs
341     .sp
342     In Perl, the sequences \el, \eL, \eu, and \eU are recognized by its string
343     handler and used to modify the case of following characters. By default, PCRE
344     does not support these escape sequences. However, if the PCRE_JAVASCRIPT_COMPAT
345     option is set, \eU matches a "U" character, and \eu can be used to define a
346     character by code point, as described in the previous section.
347     .
348     .
349 nigel 93 .SS "Absolute and relative back references"
350     .rs
351     .sp
352 ph10 208 The sequence \eg followed by an unsigned or a negative number, optionally
353     enclosed in braces, is an absolute or relative back reference. A named back
354     reference can be coded as \eg{name}. Back references are discussed
355 nigel 93 .\" HTML <a href="#backreferences">
356     .\" </a>
357     later,
358     .\"
359     following the discussion of
360     .\" HTML <a href="#subpattern">
361     .\" </a>
362     parenthesized subpatterns.
363     .\"
364     .
365     .
366 ph10 333 .SS "Absolute and relative subroutine calls"
367     .rs
368     .sp
369 ph10 345 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
370     a number enclosed either in angle brackets or single quotes, is an alternative
371     syntax for referencing a subpattern as a "subroutine". Details are discussed
372 ph10 333 .\" HTML <a href="#onigurumasubroutines">
373     .\" </a>
374     later.
375     .\"
376 ph10 345 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
377 ph10 461 synonymous. The former is a back reference; the latter is a
378 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
379     .\" </a>
380     subroutine
381     .\"
382     call.
383 ph10 333 .
384     .
385 ph10 518 .\" HTML <a name="genericchartypes"></a>
386 nigel 75 .SS "Generic character types"
387     .rs
388     .sp
389 ph10 514 Another use of backslash is for specifying generic character types:
390 nigel 75 .sp
391 ph10 182 \ed any decimal digit
392 nigel 75 \eD any character that is not a decimal digit
393 ph10 178 \eh any horizontal whitespace character
394 ph10 182 \eH any character that is not a horizontal whitespace character
395 nigel 75 \es any whitespace character
396     \eS any character that is not a whitespace character
397 ph10 178 \ev any vertical whitespace character
398 ph10 182 \eV any character that is not a vertical whitespace character
399 nigel 75 \ew any "word" character
400     \eW any "non-word" character
401     .sp
402 ph10 535 There is also the single sequence \eN, which matches a non-newline character.
403     This is the same as
404 ph10 514 .\" HTML <a href="#fullstopdot">
405     .\" </a>
406 ph10 535 the "." metacharacter
407 ph10 514 .\"
408 ph10 745 when PCRE_DOTALL is not set. Perl also uses \eN to match characters by name;
409     PCRE does not support this.
410 nigel 75 .P
411 ph10 514 Each pair of lower and upper case escape sequences partitions the complete set
412     of characters into two disjoint sets. Any given character matches one, and only
413 ph10 518 one, of each pair. The sequences can appear both inside and outside character
414 nigel 75 classes. They each match one character of the appropriate type. If the current
415 ph10 518 matching point is at the end of the subject string, all of them fail, because
416 nigel 75 there is no character to match.
417     .P
418     For compatibility with Perl, \es does not match the VT character (code 11).
419     This makes it different from the the POSIX "space" class. The \es characters
420 ph10 178 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
421 nigel 91 included in a Perl script, \es may match the VT character. In PCRE, it never
422 ph10 178 does.
423 nigel 75 .P
424 ph10 518 A "word" character is an underscore or any character that is a letter or digit.
425     By default, the definition of letters and digits is controlled by PCRE's
426     low-valued character tables, and may vary if locale-specific matching is taking
427     place (see
428     .\" HTML <a href="pcreapi.html#localesupport">
429     .\" </a>
430     "Locale support"
431     .\"
432     in the
433     .\" HREF
434     \fBpcreapi\fP
435     .\"
436     page). For example, in a French locale such as "fr_FR" in Unix-like systems,
437     or "french" in Windows, some character codes greater than 128 are used for
438     accented letters, and these are then matched by \ew. The use of locales with
439     Unicode is discouraged.
440 ph10 178 .P
441 ph10 518 By default, in UTF-8 mode, characters with values greater than 128 never match
442     \ed, \es, or \ew, and always match \eD, \eS, and \eW. These sequences retain
443     their original meanings from before UTF-8 support was available, mainly for
444 ph10 535 efficiency reasons. However, if PCRE is compiled with Unicode property support,
445 ph10 518 and the PCRE_UCP option is set, the behaviour is changed so that Unicode
446     properties are used to determine character types, as follows:
447     .sp
448     \ed any character that \ep{Nd} matches (decimal digit)
449     \es any character that \ep{Z} matches, plus HT, LF, FF, CR
450     \ew any character that \ep{L} or \ep{N} matches, plus underscore
451     .sp
452     The upper case escapes match the inverse sets of characters. Note that \ed
453 ph10 535 matches only decimal digits, whereas \ew matches any Unicode digit, as well as
454 ph10 518 any Unicode letter, and underscore. Note also that PCRE_UCP affects \eb, and
455     \eB because they are defined in terms of \ew and \eW. Matching these sequences
456     is noticeably slower when PCRE_UCP is set.
457     .P
458 ph10 579 The sequences \eh, \eH, \ev, and \eV are features that were added to Perl at
459 ph10 572 release 5.10. In contrast to the other sequences, which match only ASCII
460     characters by default, these always match certain high-valued codepoints in
461     UTF-8 mode, whether or not PCRE_UCP is set. The horizontal space characters
462     are:
463 ph10 178 .sp
464     U+0009 Horizontal tab
465     U+0020 Space
466     U+00A0 Non-break space
467     U+1680 Ogham space mark
468     U+180E Mongolian vowel separator
469     U+2000 En quad
470     U+2001 Em quad
471     U+2002 En space
472     U+2003 Em space
473     U+2004 Three-per-em space
474     U+2005 Four-per-em space
475     U+2006 Six-per-em space
476     U+2007 Figure space
477     U+2008 Punctuation space
478     U+2009 Thin space
479     U+200A Hair space
480     U+202F Narrow no-break space
481     U+205F Medium mathematical space
482     U+3000 Ideographic space
483     .sp
484     The vertical space characters are:
485     .sp
486     U+000A Linefeed
487     U+000B Vertical tab
488     U+000C Formfeed
489     U+000D Carriage return
490     U+0085 Next line
491     U+2028 Line separator
492     U+2029 Paragraph separator
493 nigel 75 .
494     .
495 ph10 231 .\" HTML <a name="newlineseq"></a>
496 nigel 93 .SS "Newline sequences"
497     .rs
498     .sp
499 ph10 231 Outside a character class, by default, the escape sequence \eR matches any
500 ph10 572 Unicode newline sequence. In non-UTF-8 mode \eR is equivalent to the following:
501 nigel 93 .sp
502     (?>\er\en|\en|\ex0b|\ef|\er|\ex85)
503     .sp
504     This is an example of an "atomic group", details of which are given
505     .\" HTML <a href="#atomicgroup">
506     .\" </a>
507     below.
508     .\"
509     This particular group matches either the two-character sequence CR followed by
510     LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
511     U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
512     line, U+0085). The two-character sequence is treated as a single unit that
513     cannot be split.
514     .P
515     In UTF-8 mode, two additional characters whose codepoints are greater than 255
516     are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
517     Unicode character property support is not needed for these characters to be
518     recognized.
519     .P
520 ph10 231 It is possible to restrict \eR to match only CR, LF, or CRLF (instead of the
521     complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
522 ph10 247 either at compile time or when the pattern is matched. (BSR is an abbrevation
523 ph10 246 for "backslash R".) This can be made the default when PCRE is built; if this is
524     the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
525     It is also possible to specify these settings by starting a pattern string with
526     one of the following sequences:
527 ph10 231 .sp
528     (*BSR_ANYCRLF) CR, LF, or CRLF only
529     (*BSR_UNICODE) any Unicode newline sequence
530     .sp
531 ph10 461 These override the default and the options given to \fBpcre_compile()\fP or
532 ph10 456 \fBpcre_compile2()\fP, but they can be overridden by options given to
533     \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP. Note that these special settings,
534     which are not Perl-compatible, are recognized only at the very start of a
535     pattern, and that they must be in upper case. If more than one of them is
536     present, the last one is used. They can be combined with a change of newline
537 ph10 518 convention; for example, a pattern can start with:
538 ph10 246 .sp
539     (*ANY)(*BSR_ANYCRLF)
540     .sp
541 ph10 518 They can also be combined with the (*UTF8) or (*UCP) special sequences. Inside
542     a character class, \eR is treated as an unrecognized escape sequence, and so
543     matches the letter "R" by default, but causes an error if PCRE_EXTRA is set.
544 nigel 93 .
545     .
546 nigel 75 .\" HTML <a name="uniextseq"></a>
547     .SS Unicode character properties
548     .rs
549     .sp
550     When PCRE is built with Unicode character property support, three additional
551 ph10 184 escape sequences that match characters with specific properties are available.
552     When not in UTF-8 mode, these sequences are of course limited to testing
553     characters whose codepoints are less than 256, but they do work in this mode.
554     The extra escape sequences are:
555 nigel 75 .sp
556 nigel 87 \ep{\fIxx\fP} a character with the \fIxx\fP property
557     \eP{\fIxx\fP} a character without the \fIxx\fP property
558     \eX an extended Unicode sequence
559 nigel 75 .sp
560 nigel 87 The property names represented by \fIxx\fP above are limited to the Unicode
561 ph10 517 script names, the general category properties, "Any", which matches any
562     character (including newline), and some special PCRE properties (described
563 ph10 535 in the
564 ph10 517 .\" HTML <a href="#extraprops">
565     .\" </a>
566 ph10 535 next section).
567 ph10 517 .\"
568     Other Perl properties such as "InMusicalSymbols" are not currently supported by
569     PCRE. Note that \eP{Any} does not match any characters, so always causes a
570     match failure.
571 nigel 75 .P
572 nigel 87 Sets of Unicode characters are defined as belonging to certain scripts. A
573     character from one of these sets can be matched using a script name. For
574     example:
575 nigel 75 .sp
576 nigel 87 \ep{Greek}
577     \eP{Han}
578     .sp
579     Those that are not part of an identified script are lumped together as
580     "Common". The current list of scripts is:
581     .P
582     Arabic,
583     Armenian,
584 ph10 491 Avestan,
585 nigel 93 Balinese,
586 ph10 491 Bamum,
587 nigel 87 Bengali,
588     Bopomofo,
589     Braille,
590     Buginese,
591     Buhid,
592     Canadian_Aboriginal,
593 ph10 491 Carian,
594     Cham,
595 nigel 87 Cherokee,
596     Common,
597     Coptic,
598 nigel 93 Cuneiform,
599 nigel 87 Cypriot,
600     Cyrillic,
601     Deseret,
602     Devanagari,
603 ph10 491 Egyptian_Hieroglyphs,
604 nigel 87 Ethiopic,
605     Georgian,
606     Glagolitic,
607     Gothic,
608     Greek,
609     Gujarati,
610     Gurmukhi,
611     Han,
612     Hangul,
613     Hanunoo,
614     Hebrew,
615     Hiragana,
616 ph10 491 Imperial_Aramaic,
617 nigel 87 Inherited,
618 ph10 491 Inscriptional_Pahlavi,
619     Inscriptional_Parthian,
620     Javanese,
621     Kaithi,
622 nigel 87 Kannada,
623     Katakana,
624 ph10 491 Kayah_Li,
625 nigel 87 Kharoshthi,
626     Khmer,
627     Lao,
628     Latin,
629 ph10 491 Lepcha,
630 nigel 87 Limbu,
631     Linear_B,
632 ph10 491 Lisu,
633     Lycian,
634     Lydian,
635 nigel 87 Malayalam,
636 ph10 491 Meetei_Mayek,
637 nigel 87 Mongolian,
638     Myanmar,
639     New_Tai_Lue,
640 nigel 93 Nko,
641 nigel 87 Ogham,
642     Old_Italic,
643     Old_Persian,
644 ph10 491 Old_South_Arabian,
645     Old_Turkic,
646     Ol_Chiki,
647 nigel 87 Oriya,
648     Osmanya,
649 nigel 93 Phags_Pa,
650     Phoenician,
651 ph10 491 Rejang,
652 nigel 87 Runic,
653 ph10 491 Samaritan,
654     Saurashtra,
655 nigel 87 Shavian,
656     Sinhala,
657 ph10 491 Sundanese,
658 nigel 87 Syloti_Nagri,
659     Syriac,
660     Tagalog,
661     Tagbanwa,
662     Tai_Le,
663 ph10 491 Tai_Tham,
664     Tai_Viet,
665 nigel 87 Tamil,
666     Telugu,
667     Thaana,
668     Thai,
669     Tibetan,
670     Tifinagh,
671     Ugaritic,
672 ph10 491 Vai,
673 nigel 87 Yi.
674     .P
675 ph10 517 Each character has exactly one Unicode general category property, specified by
676     a two-letter abbreviation. For compatibility with Perl, negation can be
677     specified by including a circumflex between the opening brace and the property
678     name. For example, \ep{^Lu} is the same as \eP{Lu}.
679 nigel 87 .P
680     If only one letter is specified with \ep or \eP, it includes all the general
681     category properties that start with that letter. In this case, in the absence
682     of negation, the curly brackets in the escape sequence are optional; these two
683     examples have the same effect:
684     .sp
685 nigel 75 \ep{L}
686     \epL
687     .sp
688 nigel 87 The following general category property codes are supported:
689 nigel 75 .sp
690     C Other
691     Cc Control
692     Cf Format
693     Cn Unassigned
694     Co Private use
695     Cs Surrogate
696     .sp
697     L Letter
698     Ll Lower case letter
699     Lm Modifier letter
700     Lo Other letter
701     Lt Title case letter
702     Lu Upper case letter
703     .sp
704     M Mark
705     Mc Spacing mark
706     Me Enclosing mark
707     Mn Non-spacing mark
708     .sp
709     N Number
710     Nd Decimal number
711     Nl Letter number
712     No Other number
713     .sp
714     P Punctuation
715     Pc Connector punctuation
716     Pd Dash punctuation
717     Pe Close punctuation
718     Pf Final punctuation
719     Pi Initial punctuation
720     Po Other punctuation
721     Ps Open punctuation
722     .sp
723     S Symbol
724     Sc Currency symbol
725     Sk Modifier symbol
726     Sm Mathematical symbol
727     So Other symbol
728     .sp
729     Z Separator
730     Zl Line separator
731     Zp Paragraph separator
732     Zs Space separator
733     .sp
734 nigel 87 The special property L& is also supported: it matches a character that has
735     the Lu, Ll, or Lt property, in other words, a letter that is not classified as
736     a modifier or "other".
737 nigel 75 .P
738 ph10 211 The Cs (Surrogate) property applies only to characters in the range U+D800 to
739     U+DFFF. Such characters are not valid in UTF-8 strings (see RFC 3629) and so
740     cannot be tested by PCRE, unless UTF-8 validity checking has been turned off
741     (see the discussion of PCRE_NO_UTF8_CHECK in the
742     .\" HREF
743     \fBpcreapi\fP
744     .\"
745 ph10 451 page). Perl does not support the Cs property.
746 ph10 211 .P
747 ph10 451 The long synonyms for property names that Perl supports (such as \ep{Letter})
748 nigel 91 are not supported by PCRE, nor is it permitted to prefix any of these
749 nigel 87 properties with "Is".
750     .P
751     No character that is in the Unicode table has the Cn (unassigned) property.
752     Instead, this property is assumed for any code point that is not in the
753     Unicode table.
754     .P
755 nigel 75 Specifying caseless matching does not affect these escape sequences. For
756     example, \ep{Lu} always matches only upper case letters.
757     .P
758     The \eX escape matches any number of Unicode characters that form an extended
759     Unicode sequence. \eX is equivalent to
760     .sp
761     (?>\ePM\epM*)
762     .sp
763     That is, it matches a character without the "mark" property, followed by zero
764     or more characters with the "mark" property, and treats the sequence as an
765     atomic group
766     .\" HTML <a href="#atomicgroup">
767     .\" </a>
768     (see below).
769     .\"
770     Characters with the "mark" property are typically accents that affect the
771 ph10 185 preceding character. None of them have codepoints less than 256, so in
772 ph10 184 non-UTF-8 mode \eX matches any one character.
773 nigel 75 .P
774 ph10 654 Note that recent versions of Perl have changed \eX to match what Unicode calls
775 ph10 628 an "extended grapheme cluster", which has a more complicated definition.
776     .P
777 nigel 75 Matching characters by Unicode property is not fast, because PCRE has to search
778     a structure that contains data for over fifteen thousand characters. That is
779     why the traditional escape sequences such as \ed and \ew do not use Unicode
780 ph10 535 properties in PCRE by default, though you can make them do so by setting the
781 ph10 518 PCRE_UCP option for \fBpcre_compile()\fP or by starting the pattern with
782     (*UCP).
783 nigel 75 .
784     .
785 ph10 517 .\" HTML <a name="extraprops"></a>
786     .SS PCRE's additional properties
787     .rs
788     .sp
789 ph10 535 As well as the standard Unicode properties described in the previous
790     section, PCRE supports four more that make it possible to convert traditional
791 ph10 517 escape sequences such as \ew and \es and POSIX character classes to use Unicode
792 ph10 518 properties. PCRE uses these non-standard, non-Perl properties internally when
793     PCRE_UCP is set. They are:
794 ph10 517 .sp
795     Xan Any alphanumeric character
796     Xps Any POSIX space character
797     Xsp Any Perl space character
798     Xwd Any Perl "word" character
799     .sp
800 ph10 535 Xan matches characters that have either the L (letter) or the N (number)
801     property. Xps matches the characters tab, linefeed, vertical tab, formfeed, or
802 ph10 517 carriage return, and any other character that has the Z (separator) property.
803 ph10 535 Xsp is the same as Xps, except that vertical tab is excluded. Xwd matches the
804 ph10 517 same characters as Xan, plus underscore.
805     .
806     .
807 ph10 168 .\" HTML <a name="resetmatchstart"></a>
808     .SS "Resetting the match start"
809     .rs
810     .sp
811 ph10 572 The escape sequence \eK causes any previously matched characters not to be
812     included in the final matched sequence. For example, the pattern:
813 ph10 168 .sp
814     foo\eKbar
815     .sp
816 ph10 172 matches "foobar", but reports that it has matched "bar". This feature is
817 ph10 168 similar to a lookbehind assertion
818     .\" HTML <a href="#lookbehind">
819     .\" </a>
820     (described below).
821     .\"
822 ph10 172 However, in this case, the part of the subject before the real match does not
823     have to be of fixed length, as lookbehind assertions do. The use of \eK does
824 ph10 168 not interfere with the setting of
825     .\" HTML <a href="#subpattern">
826     .\" </a>
827     captured substrings.
828 ph10 172 .\"
829 ph10 168 For example, when the pattern
830     .sp
831     (foo)\eKbar
832     .sp
833 ph10 172 matches "foobar", the first substring is still set to "foo".
834 ph10 500 .P
835 ph10 507 Perl documents that the use of \eK within assertions is "not well defined". In
836     PCRE, \eK is acted upon when it occurs inside positive assertions, but is
837 ph10 500 ignored in negative assertions.
838 ph10 168 .
839     .
840 nigel 75 .\" HTML <a name="smallassertions"></a>
841     .SS "Simple assertions"
842     .rs
843     .sp
844 nigel 93 The final use of backslash is for certain simple assertions. An assertion
845 nigel 63 specifies a condition that has to be met at a particular point in a match,
846     without consuming any characters from the subject string. The use of
847 nigel 75 subpatterns for more complicated assertions is described
848     .\" HTML <a href="#bigassertions">
849     .\" </a>
850     below.
851     .\"
852 nigel 91 The backslashed assertions are:
853 nigel 75 .sp
854     \eb matches at a word boundary
855     \eB matches when not at a word boundary
856 nigel 93 \eA matches at the start of the subject
857     \eZ matches at the end of the subject
858     also matches before a newline at the end of the subject
859     \ez matches only at the end of the subject
860     \eG matches at the first matching position in the subject
861 nigel 75 .sp
862 ph10 513 Inside a character class, \eb has a different meaning; it matches the backspace
863 ph10 535 character. If any other of these assertions appears in a character class, by
864 ph10 513 default it matches the corresponding literal character (for example, \eB
865     matches the letter B). However, if the PCRE_EXTRA option is set, an "invalid
866     escape sequence" error is generated instead.
867 nigel 75 .P
868 nigel 63 A word boundary is a position in the subject string where the current character
869 nigel 75 and the previous character do not both match \ew or \eW (i.e. one matches
870     \ew and the other matches \eW), or the start or end of the string if the
871 ph10 518 first or last character matches \ew, respectively. In UTF-8 mode, the meanings
872     of \ew and \eW can be changed by setting the PCRE_UCP option. When this is
873     done, it also affects \eb and \eB. Neither PCRE nor Perl has a separate "start
874     of word" or "end of word" metasequence. However, whatever follows \eb normally
875     determines which it is. For example, the fragment \eba matches "a" at the start
876     of a word.
877 nigel 75 .P
878     The \eA, \eZ, and \ez assertions differ from the traditional circumflex and
879     dollar (described in the next section) in that they only ever match at the very
880     start and end of the subject string, whatever options are set. Thus, they are
881     independent of multiline mode. These three assertions are not affected by the
882     PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
883     circumflex and dollar metacharacters. However, if the \fIstartoffset\fP
884     argument of \fBpcre_exec()\fP is non-zero, indicating that matching is to start
885     at a point other than the beginning of the subject, \eA can never match. The
886 nigel 91 difference between \eZ and \ez is that \eZ matches before a newline at the end
887     of the string as well as at the very end, whereas \ez matches only at the end.
888 nigel 75 .P
889     The \eG assertion is true only when the current matching position is at the
890     start point of the match, as specified by the \fIstartoffset\fP argument of
891     \fBpcre_exec()\fP. It differs from \eA when the value of \fIstartoffset\fP is
892     non-zero. By calling \fBpcre_exec()\fP multiple times with appropriate
893 nigel 63 arguments, you can mimic Perl's /g option, and it is in this kind of
894 nigel 75 implementation where \eG can be useful.
895     .P
896     Note, however, that PCRE's interpretation of \eG, as the start of the current
897 nigel 63 match, is subtly different from Perl's, which defines it as the end of the
898     previous match. In Perl, these can be different when the previously matched
899     string was empty. Because PCRE does just one match at a time, it cannot
900     reproduce this behaviour.
901 nigel 75 .P
902     If all the alternatives of a pattern begin with \eG, the expression is anchored
903 nigel 63 to the starting match position, and the "anchored" flag is set in the compiled
904     regular expression.
905 nigel 75 .
906     .
908 nigel 63 .rs
909     .sp
910     Outside a character class, in the default matching mode, the circumflex
911 nigel 75 character is an assertion that is true only if the current matching point is
912     at the start of the subject string. If the \fIstartoffset\fP argument of
913     \fBpcre_exec()\fP is non-zero, circumflex can never match if the PCRE_MULTILINE
914 nigel 63 option is unset. Inside a character class, circumflex has an entirely different
915 nigel 75 meaning
916     .\" HTML <a href="#characterclass">
917     .\" </a>
918     (see below).
919     .\"
920     .P
921 nigel 63 Circumflex need not be the first character of the pattern if a number of
922     alternatives are involved, but it should be the first thing in each alternative
923     in which it appears if the pattern is ever to match that branch. If all
924     possible alternatives start with a circumflex, that is, if the pattern is
925     constrained to match only at the start of the subject, it is said to be an
926     "anchored" pattern. (There are also other constructs that can cause a pattern
927     to be anchored.)
928 nigel 75 .P
929     A dollar character is an assertion that is true only if the current matching
930 nigel 63 point is at the end of the subject string, or immediately before a newline
931 nigel 91 at the end of the string (by default). Dollar need not be the last character of
932     the pattern if a number of alternatives are involved, but it should be the last
933     item in any branch in which it appears. Dollar has no special meaning in a
934     character class.
935 nigel 75 .P
936 nigel 63 The meaning of dollar can be changed so that it matches only at the very end of
937     the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
938 nigel 75 does not affect the \eZ assertion.
939     .P
940 nigel 63 The meanings of the circumflex and dollar characters are changed if the
941 nigel 91 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
942     immediately after internal newlines as well as at the start of the subject
943     string. It does not match after a newline that ends the string. A dollar
944     matches before any newlines in the string, as well as at the very end, when
945     PCRE_MULTILINE is set. When newline is specified as the two-character
946     sequence CRLF, isolated CR and LF characters do not indicate newlines.
947 nigel 75 .P
948 nigel 91 For example, the pattern /^abc$/ matches the subject string "def\enabc" (where
949     \en represents a newline) in multiline mode, but not otherwise. Consequently,
950     patterns that are anchored in single line mode because all branches start with
951     ^ are not anchored in multiline mode, and a match for circumflex is possible
952     when the \fIstartoffset\fP argument of \fBpcre_exec()\fP is non-zero. The
953     PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
954     .P
955 nigel 75 Note that the sequences \eA, \eZ, and \ez can be used to match the start and
956 nigel 63 end of the subject in both modes, and if all branches of a pattern start with
957 nigel 91 \eA it is always anchored, whether or not PCRE_MULTILINE is set.
958 nigel 75 .
959     .
960 ph10 514 .\" HTML <a name="fullstopdot"></a>
962 nigel 63 .rs
963     .sp
964     Outside a character class, a dot in the pattern matches any one character in
965 nigel 91 the subject string except (by default) a character that signifies the end of a
966 nigel 93 line. In UTF-8 mode, the matched character may be more than one byte long.
967 nigel 91 .P
968 nigel 93 When a line ending is defined as a single character, dot never matches that
969     character; when the two-character sequence CRLF is used, dot does not match CR
970     if it is immediately followed by LF, but otherwise it matches all characters
971     (including isolated CRs and LFs). When any Unicode line endings are being
972     recognized, dot does not match CR or LF or any of the other line ending
973     characters.
974     .P
975 nigel 91 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
976 nigel 93 option is set, a dot matches any one character, without exception. If the
977     two-character sequence CRLF is present in the subject string, it takes two dots
978     to match it.
979 nigel 91 .P
980     The handling of dot is entirely independent of the handling of circumflex and
981     dollar, the only relationship being that they both involve newlines. Dot has no
982     special meaning in a character class.
983 ph10 514 .P
984 ph10 579 The escape sequence \eN behaves like a dot, except that it is not affected by
985 ph10 572 the PCRE_DOTALL option. In other words, it matches any character except one
986 ph10 745 that signifies the end of a line. Perl also uses \eN to match characters by
987     name; PCRE does not support this.
988 nigel 75 .
989     .
991 nigel 63 .rs
992     .sp
993 nigel 75 Outside a character class, the escape sequence \eC matches any one byte, both
994 ph10 737 in and out of UTF-8 mode. Unlike a dot, it always matches line-ending
995 nigel 93 characters. The feature is provided in Perl in order to match individual bytes
996 ph10 737 in UTF-8 mode, but it is unclear how it can usefully be used. Because \eC
997     breaks up characters into individual bytes, matching one byte with \eC in UTF-8
998     mode means that the rest of the string may start with a malformed UTF-8
999     character. This has undefined results, because PCRE assumes that it is dealing
1000     with valid UTF-8 strings (and by default it checks this at the start of
1001     processing unless the PCRE_NO_UTF8_CHECK option is used).
1002 nigel 75 .P
1003     PCRE does not allow \eC to appear in lookbehind assertions
1004     .\" HTML <a href="#lookbehind">
1005     .\" </a>
1006     (described below),
1007     .\"
1008     because in UTF-8 mode this would make it impossible to calculate the length of
1009     the lookbehind.
1010 ph10 737 .P
1011     In general, the \eC escape sequence is best avoided in UTF-8 mode. However, one
1012 ph10 738 way of using it that avoids the problem of malformed UTF-8 characters is to
1013     use a lookahead to check the length of the next character, as in this pattern
1014 ph10 737 (ignore white space and line breaks):
1015     .sp
1016     (?| (?=[\ex00-\ex7f])(\eC) |
1017 ph10 738 (?=[\ex80-\ex{7ff}])(\eC)(\eC) |
1018     (?=[\ex{800}-\ex{ffff}])(\eC)(\eC)(\eC) |
1019 ph10 737 (?=[\ex{10000}-\ex{1fffff}])(\eC)(\eC)(\eC)(\eC))
1020     .sp
1021 ph10 738 A group that starts with (?| resets the capturing parentheses numbers in each
1022     alternative (see
1023 ph10 737 .\" HTML <a href="#dupsubpatternnumber">
1024     .\" </a>
1025     "Duplicate Subpattern Numbers"
1026     .\"
1027 ph10 738 below). The assertions at the start of each branch check the next UTF-8
1028     character for values whose encoding uses 1, 2, 3, or 4 bytes, respectively. The
1029 ph10 737 character's individual bytes are then captured by the appropriate number of
1030     groups.
1031 nigel 75 .
1032     .
1033     .\" HTML <a name="characterclass"></a>
1035 nigel 63 .rs
1036     .sp
1037     An opening square bracket introduces a character class, terminated by a closing
1038 ph10 461 square bracket. A closing square bracket on its own is not special by default.
1039     However, if the PCRE_JAVASCRIPT_COMPAT option is set, a lone closing square
1040 ph10 456 bracket causes a compile-time error. If a closing square bracket is required as
1041     a member of the class, it should be the first data character in the class
1042     (after an initial circumflex, if present) or escaped with a backslash.
1043 nigel 75 .P
1044 nigel 63 A character class matches a single character in the subject. In UTF-8 mode, the
1045 ph10 456 character may be more than one byte long. A matched character must be in the
1046     set of characters defined by the class, unless the first character in the class
1047 nigel 63 definition is a circumflex, in which case the subject character must not be in
1048     the set defined by the class. If a circumflex is actually required as a member
1049     of the class, ensure it is not the first character, or escape it with a
1050     backslash.
1051 nigel 75 .P
1052 nigel 63 For example, the character class [aeiou] matches any lower case vowel, while
1053     [^aeiou] matches any character that is not a lower case vowel. Note that a
1054 nigel 75 circumflex is just a convenient notation for specifying the characters that
1055     are in the class by enumerating those that are not. A class that starts with a
1056 ph10 456 circumflex is not an assertion; it still consumes a character from the subject
1057 nigel 75 string, and therefore it fails if the current pointer is at the end of the
1058     string.
1059     .P
1060 nigel 63 In UTF-8 mode, characters with values greater than 255 can be included in a
1061 nigel 75 class as a literal string of bytes, or by using the \ex{ escaping mechanism.
1062     .P
1063 nigel 63 When caseless matching is set, any letters in a class represent both their
1064     upper case and lower case versions, so for example, a caseless [aeiou] matches
1065     "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
1066 nigel 77 caseful version would. In UTF-8 mode, PCRE always understands the concept of
1067     case for characters whose values are less than 128, so caseless matching is
1068     always possible. For characters with higher values, the concept of case is
1069     supported if PCRE is compiled with Unicode property support, but not otherwise.
1070 ph10 456 If you want to use caseless matching in UTF8-mode for characters 128 and above,
1071     you must ensure that PCRE is compiled with Unicode property support as well as
1072     with UTF-8 support.
1073 nigel 75 .P
1074 nigel 93 Characters that might indicate line breaks are never treated in any special way
1075     when matching character classes, whatever line-ending sequence is in use, and
1076     whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
1077     such as [^a] always matches one of these characters.
1078 nigel 75 .P
1079 nigel 63 The minus (hyphen) character can be used to specify a range of characters in a
1080     character class. For example, [d-m] matches any letter between d and m,
1081     inclusive. If a minus character is required in a class, it must be escaped with
1082     a backslash or appear in a position where it cannot be interpreted as
1083     indicating a range, typically as the first or last character in the class.
1084 nigel 75 .P
1085 nigel 63 It is not possible to have the literal character "]" as the end character of a
1086     range. A pattern such as [W-]46] is interpreted as a class of two characters
1087     ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
1088     "-46]". However, if the "]" is escaped with a backslash it is interpreted as
1089 nigel 75 the end of range, so [W-\e]46] is interpreted as a class containing a range
1090     followed by two other characters. The octal or hexadecimal representation of
1091     "]" can also be used to end a range.
1092     .P
1093 nigel 63 Ranges operate in the collating sequence of character values. They can also be
1094 nigel 75 used for characters specified numerically, for example [\e000-\e037]. In UTF-8
1095 nigel 63 mode, ranges can include characters whose values are greater than 255, for
1096 nigel 75 example [\ex{100}-\ex{2ff}].
1097     .P
1098 nigel 63 If a range that includes letters is used when caseless matching is set, it
1099     matches the letters in either case. For example, [W-c] is equivalent to
1100 nigel 75 [][\e\e^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
1101 ph10 139 tables for a French locale are in use, [\exc8-\excb] matches accented E
1102 nigel 75 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
1103     characters with values greater than 128 only when it is compiled with Unicode
1104     property support.
1105     .P
1106 ph10 575 The character escape sequences \ed, \eD, \eh, \eH, \ep, \eP, \es, \eS, \ev,
1107     \eV, \ew, and \eW may appear in a character class, and add the characters that
1108     they match to the class. For example, [\edABCDEF] matches any hexadecimal
1109     digit. In UTF-8 mode, the PCRE_UCP option affects the meanings of \ed, \es, \ew
1110     and their upper case partners, just as it does when they appear outside a
1111     character class, as described in the section entitled
1112     .\" HTML <a href="#genericchartypes">
1113     .\" </a>
1114     "Generic character types"
1115     .\"
1116     above. The escape sequence \eb has a different meaning inside a character
1117     class; it matches the backspace character. The sequences \eB, \eN, \eR, and \eX
1118     are not special inside a character class. Like any other unrecognized escape
1119     sequences, they are treated as the literal characters "B", "N", "R", and "X" by
1120     default, but cause an error if the PCRE_EXTRA option is set.
1121     .P
1122     A circumflex can conveniently be used with the upper case character types to
1123 ph10 518 specify a more restricted set of characters than the matching lower case type.
1124 ph10 575 For example, the class [^\eW_] matches any letter or digit, but not underscore,
1125 ph10 579 whereas [\ew] includes underscore. A positive character class should be read as
1126     "something OR something OR ..." and a negative class as "NOT something AND NOT
1127 ph10 575 something AND NOT ...".
1128 nigel 75 .P
1129     The only metacharacters that are recognized in character classes are backslash,
1130     hyphen (only where it can be interpreted as specifying a range), circumflex
1131     (only at the start), opening square bracket (only when it can be interpreted as
1132     introducing a POSIX class name - see the next section), and the terminating
1133     closing square bracket. However, escaping other non-alphanumeric characters
1134     does no harm.
1135     .
1136     .
1138 nigel 63 .rs
1139     .sp
1140 nigel 75 Perl supports the POSIX notation for character classes. This uses names
1141 nigel 63 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
1142     this notation. For example,
1143 nigel 75 .sp
1144 nigel 63 [01[:alpha:]%]
1145 nigel 75 .sp
1146 nigel 63 matches "0", "1", any alphabetic character, or "%". The supported class names
1147 ph10 518 are:
1148 nigel 75 .sp
1149 nigel 63 alnum letters and digits
1150     alpha letters
1151     ascii character codes 0 - 127
1152     blank space or tab only
1153     cntrl control characters
1154 nigel 75 digit decimal digits (same as \ed)
1155 nigel 63 graph printing characters, excluding space
1156     lower lower case letters
1157     print printing characters, including space
1158 ph10 518 punct printing characters, excluding letters and digits and space
1159 nigel 75 space white space (not quite the same as \es)
1160 nigel 63 upper upper case letters
1161 nigel 75 word "word" characters (same as \ew)
1162 nigel 63 xdigit hexadecimal digits
1163 nigel 75 .sp
1164 nigel 63 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
1165     space (32). Notice that this list includes the VT character (code 11). This
1166 nigel 75 makes "space" different to \es, which does not include VT (for Perl
1167 nigel 63 compatibility).
1168 nigel 75 .P
1169 nigel 63 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
1170     5.8. Another Perl extension is negation, which is indicated by a ^ character
1171     after the colon. For example,
1172 nigel 75 .sp
1173 nigel 63 [12[:^digit:]]
1174 nigel 75 .sp
1175 nigel 63 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
1176     syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
1177     supported, and an error is given if they are encountered.
1178 nigel 75 .P
1179 ph10 518 By default, in UTF-8 mode, characters with values greater than 128 do not match
1180     any of the POSIX character classes. However, if the PCRE_UCP option is passed
1181 ph10 535 to \fBpcre_compile()\fP, some of the classes are changed so that Unicode
1182     character properties are used. This is achieved by replacing the POSIX classes
1183 ph10 518 by other sequences, as follows:
1184     .sp
1185     [:alnum:] becomes \ep{Xan}
1186     [:alpha:] becomes \ep{L}
1187 ph10 535 [:blank:] becomes \eh
1188 ph10 518 [:digit:] becomes \ep{Nd}
1189     [:lower:] becomes \ep{Ll}
1190 ph10 535 [:space:] becomes \ep{Xps}
1191 ph10 518 [:upper:] becomes \ep{Lu}
1192     [:word:] becomes \ep{Xwd}
1193     .sp
1194     Negated versions, such as [:^alpha:] use \eP instead of \ep. The other POSIX
1195     classes are unchanged, and match only characters with code points less than
1196     128.
1197 nigel 75 .
1198     .
1199     .SH "VERTICAL BAR"
1200 nigel 63 .rs
1201     .sp
1202     Vertical bar characters are used to separate alternative patterns. For example,
1203     the pattern
1204 nigel 75 .sp
1205 nigel 63 gilbert|sullivan
1206 nigel 75 .sp
1207 nigel 63 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
1208 nigel 91 and an empty alternative is permitted (matching the empty string). The matching
1209     process tries each alternative in turn, from left to right, and the first one
1210     that succeeds is used. If the alternatives are within a subpattern
1211 nigel 75 .\" HTML <a href="#subpattern">
1212     .\" </a>
1213     (defined below),
1214     .\"
1215     "succeeds" means matching the rest of the main pattern as well as the
1216     alternative in the subpattern.
1217     .
1218     .
1220 nigel 63 .rs
1221     .sp
1222     The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1223 ph10 231 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1224     the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1225     The option letters are
1226 nigel 75 .sp
1227 nigel 63 i for PCRE_CASELESS
1228     m for PCRE_MULTILINE
1229     s for PCRE_DOTALL
1230     x for PCRE_EXTENDED
1231 nigel 75 .sp
1232 nigel 63 For example, (?im) sets caseless, multiline matching. It is also possible to
1233     unset these options by preceding the letter with a hyphen, and a combined
1234     setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1235     PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1236     permitted. If a letter appears both before and after the hyphen, the option is
1237     unset.
1238 nigel 75 .P
1239 ph10 231 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1240     changed in the same way as the Perl-compatible options by using the characters
1241     J, U and X respectively.
1242     .P
1243 ph10 412 When one of these option changes occurs at top level (that is, not inside
1244     subpattern parentheses), the change applies to the remainder of the pattern
1245     that follows. If the change is placed right at the start of a pattern, PCRE
1246     extracts it into the global options (and it will therefore show up in data
1247     extracted by the \fBpcre_fullinfo()\fP function).
1248 nigel 75 .P
1249 nigel 93 An option change within a subpattern (see below for a description of
1250 ph10 572 subpatterns) affects only that part of the subpattern that follows it, so
1251 nigel 75 .sp
1252 nigel 63 (a(?i)b)c
1253 nigel 75 .sp
1254 nigel 63 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1255     By this means, options can be made to have different settings in different
1256     parts of the pattern. Any changes made in one alternative do carry on
1257     into subsequent branches within the same subpattern. For example,
1258 nigel 75 .sp
1259 nigel 63 (a(?i)b|c)
1260 nigel 75 .sp
1261 nigel 63 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1262     branch is abandoned before the option setting. This is because the effects of
1263     option settings happen at compile time. There would be some very weird
1264     behaviour otherwise.
1265 ph10 251 .P
1266     \fBNote:\fP There are other PCRE-specific options that can be set by the
1267     application when the compile or match functions are called. In some cases the
1268 ph10 412 pattern can contain special leading sequences such as (*CRLF) to override what
1269     the application has set or what has been defaulted. Details are given in the
1270     section entitled
1271 ph10 251 .\" HTML <a href="#newlineseq">
1272     .\" </a>
1273     "Newline sequences"
1274     .\"
1275 ph10 518 above. There are also the (*UTF8) and (*UCP) leading sequences that can be used
1276     to set UTF-8 and Unicode property modes; they are equivalent to setting the
1277     PCRE_UTF8 and the PCRE_UCP options, respectively.
1278 nigel 75 .
1279     .
1280     .\" HTML <a name="subpattern"></a>
1281 nigel 63 .SH SUBPATTERNS
1282     .rs
1283     .sp
1284     Subpatterns are delimited by parentheses (round brackets), which can be nested.
1285 nigel 75 Turning part of a pattern into a subpattern does two things:
1286     .sp
1287 nigel 63 1. It localizes a set of alternatives. For example, the pattern
1288 nigel 75 .sp
1289 nigel 63 cat(aract|erpillar|)
1290 nigel 75 .sp
1291 ph10 572 matches "cataract", "caterpillar", or "cat". Without the parentheses, it would
1292     match "cataract", "erpillar" or an empty string.
1293 nigel 75 .sp
1294     2. It sets up the subpattern as a capturing subpattern. This means that, when
1295     the whole pattern matches, that portion of the subject string that matched the
1296     subpattern is passed back to the caller via the \fIovector\fP argument of
1297     \fBpcre_exec()\fP. Opening parentheses are counted from left to right (starting
1298 ph10 572 from 1) to obtain numbers for the capturing subpatterns. For example, if the
1299     string "the red king" is matched against the pattern
1300 nigel 75 .sp
1301 nigel 63 the ((red|white) (king|queen))
1302 nigel 75 .sp
1303 nigel 63 the captured substrings are "red king", "red", and "king", and are numbered 1,
1304     2, and 3, respectively.
1305 nigel 75 .P
1306 nigel 63 The fact that plain parentheses fulfil two functions is not always helpful.
1307     There are often times when a grouping subpattern is required without a
1308     capturing requirement. If an opening parenthesis is followed by a question mark
1309     and a colon, the subpattern does not do any capturing, and is not counted when
1310     computing the number of any subsequent capturing subpatterns. For example, if
1311     the string "the white queen" is matched against the pattern
1312 nigel 75 .sp
1313 nigel 63 the ((?:red|white) (king|queen))
1314 nigel 75 .sp
1315 nigel 63 the captured substrings are "white queen" and "queen", and are numbered 1 and
1316 nigel 93 2. The maximum number of capturing subpatterns is 65535.
1317 nigel 75 .P
1318 nigel 63 As a convenient shorthand, if any option settings are required at the start of
1319     a non-capturing subpattern, the option letters may appear between the "?" and
1320     the ":". Thus the two patterns
1321 nigel 75 .sp
1322 nigel 63 (?i:saturday|sunday)
1323     (?:(?i)saturday|sunday)
1324 nigel 75 .sp
1325 nigel 63 match exactly the same set of strings. Because alternative branches are tried
1326     from left to right, and options are not reset until the end of the subpattern
1327     is reached, an option setting in one branch does affect subsequent branches, so
1328     the above patterns match "SUNDAY" as well as "Saturday".
1329 nigel 75 .
1330     .
1331 ph10 456 .\" HTML <a name="dupsubpatternnumber"></a>
1333     .rs
1334     .sp
1335 ph10 182 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1336     the same numbers for its capturing parentheses. Such a subpattern starts with
1337     (?| and is itself a non-capturing subpattern. For example, consider this
1338 ph10 175 pattern:
1339     .sp
1340     (?|(Sat)ur|(Sun))day
1341 ph10 182 .sp
1342     Because the two alternatives are inside a (?| group, both sets of capturing
1343     parentheses are numbered one. Thus, when the pattern matches, you can look
1344     at captured substring number one, whichever alternative matched. This construct
1345     is useful when you want to capture part, but not all, of one of a number of
1346     alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1347 ph10 175 number is reset at the start of each branch. The numbers of any capturing
1348 ph10 572 parentheses that follow the subpattern start after the highest number used in
1349     any branch. The following example is taken from the Perl documentation. The
1350     numbers underneath show in which buffer the captured content will be stored.
1351 ph10 175 .sp
1352     # before ---------------branch-reset----------- after
1353     / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1354     # 1 2 2 3 2 3 4
1355 ph10 182 .sp
1356 ph10 488 A back reference to a numbered subpattern uses the most recent value that is
1357     set for that number by any subpattern. The following pattern matches "abcabc"
1358     or "defdef":
1359 ph10 456 .sp
1360 ph10 461 /(?|(abc)|(def))\e1/
1361 ph10 456 .sp
1362 ph10 716 In contrast, a subroutine call to a numbered subpattern always refers to the
1363     first one in the pattern with the given number. The following pattern matches
1364     "abcabc" or "defabc":
1365 ph10 456 .sp
1366     /(?|(abc)|(def))(?1)/
1367     .sp
1368 ph10 459 If a
1369     .\" HTML <a href="#conditions">
1370     .\" </a>
1371     condition test
1372     .\"
1373     for a subpattern's having matched refers to a non-unique number, the test is
1374     true if any of the subpatterns of that number have matched.
1375     .P
1376     An alternative approach to using this "branch reset" feature is to use
1377 ph10 175 duplicate named subpatterns, as described in the next section.
1378     .
1379     .
1380 nigel 75 .SH "NAMED SUBPATTERNS"
1381 nigel 63 .rs
1382     .sp
1383     Identifying capturing parentheses by number is simple, but it can be very hard
1384     to keep track of the numbers in complicated regular expressions. Furthermore,
1385 nigel 75 if an expression is modified, the numbers may change. To help with this
1386 nigel 93 difficulty, PCRE supports the naming of subpatterns. This feature was not
1387     added to Perl until release 5.10. Python had the feature earlier, and PCRE
1388     introduced it at release 4.0, using the Python syntax. PCRE now supports both
1389 ph10 459 the Perl and the Python syntax. Perl allows identically numbered subpatterns to
1390     have different names, but PCRE does not.
1391 nigel 93 .P
1392     In PCRE, a subpattern can be named in one of three ways: (?<name>...) or
1393     (?'name'...) as in Perl, or (?P<name>...) as in Python. References to capturing
1394 nigel 91 parentheses from other parts of the pattern, such as
1395     .\" HTML <a href="#backreferences">
1396     .\" </a>
1397 ph10 488 back references,
1398 nigel 91 .\"
1399     .\" HTML <a href="#recursion">
1400     .\" </a>
1401     recursion,
1402     .\"
1403     and
1404     .\" HTML <a href="#conditions">
1405     .\" </a>
1406     conditions,
1407     .\"
1408     can be made by name as well as by number.
1409 nigel 75 .P
1410 nigel 91 Names consist of up to 32 alphanumeric characters and underscores. Named
1411 nigel 93 capturing parentheses are still allocated numbers as well as names, exactly as
1412     if the names were not present. The PCRE API provides function calls for
1413     extracting the name-to-number translation table from a compiled pattern. There
1414     is also a convenience function for extracting a captured substring by name.
1415 nigel 91 .P
1416     By default, a name must be unique within a pattern, but it is possible to relax
1417 ph10 457 this constraint by setting the PCRE_DUPNAMES option at compile time. (Duplicate
1418 ph10 461 names are also always permitted for subpatterns with the same number, set up as
1419 ph10 457 described in the previous section.) Duplicate names can be useful for patterns
1420     where only one instance of the named parentheses can match. Suppose you want to
1421     match the name of a weekday, either as a 3-letter abbreviation or as the full
1422     name, and in both cases you want to extract the abbreviation. This pattern
1423     (ignoring the line breaks) does the job:
1424 nigel 91 .sp
1425 nigel 93 (?<DN>Mon|Fri|Sun)(?:day)?|
1426     (?<DN>Tue)(?:sday)?|
1427     (?<DN>Wed)(?:nesday)?|
1428     (?<DN>Thu)(?:rsday)?|
1429     (?<DN>Sat)(?:urday)?
1430 nigel 91 .sp
1431     There are five capturing substrings, but only one is ever set after a match.
1432 ph10 182 (An alternative way of solving this problem is to use a "branch reset"
1433 ph10 175 subpattern, as described in the previous section.)
1434     .P
1435 nigel 91 The convenience function for extracting the data by name returns the substring
1436 nigel 93 for the first (and in this example, the only) subpattern of that name that
1437 ph10 461 matched. This saves searching to find which numbered subpattern it was.
1438 ph10 459 .P
1439 ph10 488 If you make a back reference to a non-unique named subpattern from elsewhere in
1440 ph10 459 the pattern, the one that corresponds to the first occurrence of the name is
1441     used. In the absence of duplicate numbers (see the previous section) this is
1442     the one with the lowest number. If you use a named reference in a condition
1443     test (see the
1444     .\"
1445     .\" HTML <a href="#conditions">
1446     .\" </a>
1447     section about conditions
1448     .\"
1449 ph10 461 below), either to check whether a subpattern has matched, or to check for
1450 ph10 459 recursion, all subpatterns with the same name are tested. If the condition is
1451     true for any one of them, the overall condition is true. This is the same
1452     behaviour as testing by number. For further details of the interfaces for
1453     handling named subpatterns, see the
1454 nigel 63 .\" HREF
1455 nigel 75 \fBpcreapi\fP
1456 nigel 63 .\"
1457     documentation.
1458 ph10 385 .P
1459     \fBWarning:\fP You cannot use different names to distinguish between two
1460 ph10 457 subpatterns with the same number because PCRE uses only the numbers when
1461     matching. For this reason, an error is given at compile time if different names
1462     are given to subpatterns with the same number. However, you can give the same
1463     name to subpatterns with the same number, even when PCRE_DUPNAMES is not set.
1464 nigel 75 .
1465     .
1466 nigel 63 .SH REPETITION
1467     .rs
1468     .sp
1469     Repetition is specified by quantifiers, which can follow any of the following
1470     items:
1471 nigel 75 .sp
1472 nigel 63 a literal data character
1473 nigel 93 the dot metacharacter
1474 nigel 75 the \eC escape sequence
1475     the \eX escape sequence (in UTF-8 mode with Unicode properties)
1476 nigel 93 the \eR escape sequence
1477 ph10 572 an escape such as \ed or \epL that matches a single character
1478 nigel 63 a character class
1479     a back reference (see next section)
1480 ph10 637 a parenthesized subpattern (including assertions)
1481 ph10 716 a subroutine call to a subpattern (recursive or otherwise)
1482 nigel 75 .sp
1483 nigel 63 The general repetition quantifier specifies a minimum and maximum number of
1484     permitted matches, by giving the two numbers in curly brackets (braces),
1485     separated by a comma. The numbers must be less than 65536, and the first must
1486     be less than or equal to the second. For example:
1487 nigel 75 .sp
1488 nigel 63 z{2,4}
1489 nigel 75 .sp
1490 nigel 63 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1491     character. If the second number is omitted, but the comma is present, there is
1492     no upper limit; if the second number and the comma are both omitted, the
1493     quantifier specifies an exact number of required matches. Thus
1494 nigel 75 .sp
1495 nigel 63 [aeiou]{3,}
1496 nigel 75 .sp
1497 nigel 63 matches at least 3 successive vowels, but may match many more, while
1498 nigel 75 .sp
1499     \ed{8}
1500     .sp
1501 nigel 63 matches exactly 8 digits. An opening curly bracket that appears in a position
1502     where a quantifier is not allowed, or one that does not match the syntax of a
1503     quantifier, is taken as a literal character. For example, {,6} is not a
1504     quantifier, but a literal string of four characters.
1505 nigel 75 .P
1506 nigel 63 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1507 nigel 75 bytes. Thus, for example, \ex{100}{2} matches two UTF-8 characters, each of
1508     which is represented by a two-byte sequence. Similarly, when Unicode property
1509     support is available, \eX{3} matches three Unicode extended sequences, each of
1510     which may be several bytes long (and they may be of different lengths).
1511     .P
1512 nigel 63 The quantifier {0} is permitted, causing the expression to behave as if the
1513 ph10 345 previous item and the quantifier were not present. This may be useful for
1514     subpatterns that are referenced as
1515 ph10 335 .\" HTML <a href="#subpatternsassubroutines">
1516     .\" </a>
1517     subroutines
1518     .\"
1519 ph10 572 from elsewhere in the pattern (but see also the section entitled
1520     .\" HTML <a href="#subdefine">
1521     .\" </a>
1522     "Defining subpatterns for use by reference only"
1523     .\"
1524     below). Items other than subpatterns that have a {0} quantifier are omitted
1525     from the compiled pattern.
1526 nigel 75 .P
1527 nigel 93 For convenience, the three most common quantifiers have single-character
1528     abbreviations:
1529 nigel 75 .sp
1530 nigel 63 * is equivalent to {0,}
1531     + is equivalent to {1,}
1532     ? is equivalent to {0,1}
1533 nigel 75 .sp
1534 nigel 63 It is possible to construct infinite loops by following a subpattern that can
1535     match no characters with a quantifier that has no upper limit, for example:
1536 nigel 75 .sp
1537 nigel 63 (a?)*
1538 nigel 75 .sp
1539 nigel 63 Earlier versions of Perl and PCRE used to give an error at compile time for
1540     such patterns. However, because there are cases where this can be useful, such
1541     patterns are now accepted, but if any repetition of the subpattern does in fact
1542     match no characters, the loop is forcibly broken.
1543 nigel 75 .P
1544 nigel 63 By default, the quantifiers are "greedy", that is, they match as much as
1545     possible (up to the maximum number of permitted times), without causing the
1546     rest of the pattern to fail. The classic example of where this gives problems
1547 nigel 75 is in trying to match comments in C programs. These appear between /* and */
1548     and within the comment, individual * and / characters may appear. An attempt to
1549     match C comments by applying the pattern
1550     .sp
1551     /\e*.*\e*/
1552     .sp
1553 nigel 63 to the string
1554 nigel 75 .sp
1555     /* first comment */ not comment /* second comment */
1556     .sp
1557 nigel 63 fails, because it matches the entire string owing to the greediness of the .*
1558     item.
1559 nigel 75 .P
1560 nigel 63 However, if a quantifier is followed by a question mark, it ceases to be
1561     greedy, and instead matches the minimum number of times possible, so the
1562     pattern
1563 nigel 75 .sp
1564     /\e*.*?\e*/
1565     .sp
1566 nigel 63 does the right thing with the C comments. The meaning of the various
1567     quantifiers is not otherwise changed, just the preferred number of matches.
1568     Do not confuse this use of question mark with its use as a quantifier in its
1569     own right. Because it has two uses, it can sometimes appear doubled, as in
1570 nigel 75 .sp
1571     \ed??\ed
1572     .sp
1573 nigel 63 which matches one digit by preference, but can match two if that is the only
1574     way the rest of the pattern matches.
1575 nigel 75 .P
1576 nigel 93 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1577 nigel 63 the quantifiers are not greedy by default, but individual ones can be made
1578     greedy by following them with a question mark. In other words, it inverts the
1579     default behaviour.
1580 nigel 75 .P
1581 nigel 63 When a parenthesized subpattern is quantified with a minimum repeat count that
1582 nigel 75 is greater than 1 or with a limited maximum, more memory is required for the
1583 nigel 63 compiled pattern, in proportion to the size of the minimum or maximum.
1584 nigel 75 .P
1585 nigel 63 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1586 nigel 93 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1587 nigel 63 implicitly anchored, because whatever follows will be tried against every
1588     character position in the subject string, so there is no point in retrying the
1589     overall match at any position after the first. PCRE normally treats such a
1590 nigel 75 pattern as though it were preceded by \eA.
1591     .P
1592 nigel 63 In cases where it is known that the subject string contains no newlines, it is
1593     worth setting PCRE_DOTALL in order to obtain this optimization, or
1594     alternatively using ^ to indicate anchoring explicitly.
1595 nigel 75 .P
1596 nigel 63 However, there is one situation where the optimization cannot be used. When .*
1597 ph10 488 is inside capturing parentheses that are the subject of a back reference
1598 nigel 93 elsewhere in the pattern, a match at the start may fail where a later one
1599     succeeds. Consider, for example:
1600 nigel 75 .sp
1601     (.*)abc\e1
1602     .sp
1603 nigel 63 If the subject is "xyz123abc123" the match point is the fourth character. For
1604     this reason, such a pattern is not implicitly anchored.
1605 nigel 75 .P
1606 nigel 63 When a capturing subpattern is repeated, the value captured is the substring
1607     that matched the final iteration. For example, after
1608 nigel 75 .sp
1609     (tweedle[dume]{3}\es*)+
1610     .sp
1611 nigel 63 has matched "tweedledum tweedledee" the value of the captured substring is
1612     "tweedledee". However, if there are nested capturing subpatterns, the
1613     corresponding captured values may have been set in previous iterations. For
1614     example, after
1615 nigel 75 .sp
1616 nigel 63 /(a|(b))+/
1617 nigel 75 .sp
1618 nigel 63 matches "aba" the value of the second captured substring is "b".
1619 nigel 75 .
1620     .
1621     .\" HTML <a name="atomicgroup"></a>
1623 nigel 63 .rs
1624     .sp
1625 nigel 93 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1626     repetition, failure of what follows normally causes the repeated item to be
1627     re-evaluated to see if a different number of repeats allows the rest of the
1628     pattern to match. Sometimes it is useful to prevent this, either to change the
1629     nature of the match, or to cause it fail earlier than it otherwise might, when
1630     the author of the pattern knows there is no point in carrying on.
1631 nigel 75 .P
1632     Consider, for example, the pattern \ed+foo when applied to the subject line
1633     .sp
1634 nigel 63 123456bar
1635 nigel 75 .sp
1636 nigel 63 After matching all 6 digits and then failing to match "foo", the normal
1637 nigel 75 action of the matcher is to try again with only 5 digits matching the \ed+
1638 nigel 63 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1639     (a term taken from Jeffrey Friedl's book) provides the means for specifying
1640     that once a subpattern has matched, it is not to be re-evaluated in this way.
1641 nigel 75 .P
1642 nigel 93 If we use atomic grouping for the previous example, the matcher gives up
1643 nigel 63 immediately on failing to match "foo" the first time. The notation is a kind of
1644     special parenthesis, starting with (?> as in this example:
1645 nigel 75 .sp
1646     (?>\ed+)foo
1647     .sp
1648 nigel 63 This kind of parenthesis "locks up" the part of the pattern it contains once
1649     it has matched, and a failure further into the pattern is prevented from
1650     backtracking into it. Backtracking past it to previous items, however, works as
1651     normal.
1652 nigel 75 .P
1653 nigel 63 An alternative description is that a subpattern of this type matches the string
1654     of characters that an identical standalone pattern would match, if anchored at
1655     the current point in the subject string.
1656 nigel 75 .P
1657 nigel 63 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1658     the above example can be thought of as a maximizing repeat that must swallow
1659 nigel 75 everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1660 nigel 63 number of digits they match in order to make the rest of the pattern match,
1661 nigel 75 (?>\ed+) can only match an entire sequence of digits.
1662     .P
1663 nigel 63 Atomic groups in general can of course contain arbitrarily complicated
1664     subpatterns, and can be nested. However, when the subpattern for an atomic
1665     group is just a single repeated item, as in the example above, a simpler
1666     notation, called a "possessive quantifier" can be used. This consists of an
1667     additional + character following a quantifier. Using this notation, the
1668     previous example can be rewritten as
1669 nigel 75 .sp
1670     \ed++foo
1671     .sp
1672 ph10 208 Note that a possessive quantifier can be used with an entire group, for
1673     example:
1674     .sp
1675     (abc|xyz){2,3}+
1676     .sp
1677 nigel 63 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1678     option is ignored. They are a convenient notation for the simpler forms of
1679 nigel 93 atomic group. However, there is no difference in the meaning of a possessive
1680     quantifier and the equivalent atomic group, though there may be a performance
1681     difference; possessive quantifiers should be slightly faster.
1682 nigel 75 .P
1683 nigel 93 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1684     Jeffrey Friedl originated the idea (and the name) in the first edition of his
1685     book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1686     package, and PCRE copied it from there. It ultimately found its way into Perl
1687     at release 5.10.
1688 nigel 75 .P
1689 nigel 93 PCRE has an optimization that automatically "possessifies" certain simple
1690     pattern constructs. For example, the sequence A+B is treated as A++B because
1691     there is no point in backtracking into a sequence of A's when B must follow.
1692     .P
1693 nigel 63 When a pattern contains an unlimited repeat inside a subpattern that can itself
1694     be repeated an unlimited number of times, the use of an atomic group is the
1695     only way to avoid some failing matches taking a very long time indeed. The
1696     pattern
1697 nigel 75 .sp
1698     (\eD+|<\ed+>)*[!?]
1699     .sp
1700 nigel 63 matches an unlimited number of substrings that either consist of non-digits, or
1701     digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1702     quickly. However, if it is applied to
1703 nigel 75 .sp
1704 nigel 63 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1705 nigel 75 .sp
1706 nigel 63 it takes a long time before reporting failure. This is because the string can
1707 nigel 75 be divided between the internal \eD+ repeat and the external * repeat in a
1708     large number of ways, and all have to be tried. (The example uses [!?] rather
1709     than a single character at the end, because both PCRE and Perl have an
1710     optimization that allows for fast failure when a single character is used. They
1711     remember the last single character that is required for a match, and fail early
1712     if it is not present in the string.) If the pattern is changed so that it uses
1713     an atomic group, like this:
1714     .sp
1715     ((?>\eD+)|<\ed+>)*[!?]
1716     .sp
1717 nigel 63 sequences of non-digits cannot be broken, and failure happens quickly.
1718 nigel 75 .
1719     .
1720     .\" HTML <a name="backreferences"></a>
1722 nigel 63 .rs
1723     .sp
1724     Outside a character class, a backslash followed by a digit greater than 0 (and
1725     possibly further digits) is a back reference to a capturing subpattern earlier
1726     (that is, to its left) in the pattern, provided there have been that many
1727     previous capturing left parentheses.
1728 nigel 75 .P
1729 nigel 63 However, if the decimal number following the backslash is less than 10, it is
1730     always taken as a back reference, and causes an error only if there are not
1731     that many capturing left parentheses in the entire pattern. In other words, the
1732     parentheses that are referenced need not be to the left of the reference for
1733 nigel 91 numbers less than 10. A "forward back reference" of this type can make sense
1734     when a repetition is involved and the subpattern to the right has participated
1735     in an earlier iteration.
1736     .P
1737 nigel 93 It is not possible to have a numerical "forward back reference" to a subpattern
1738     whose number is 10 or more using this syntax because a sequence such as \e50 is
1739     interpreted as a character defined in octal. See the subsection entitled
1740 nigel 91 "Non-printing characters"
1741 nigel 75 .\" HTML <a href="#digitsafterbackslash">
1742     .\" </a>
1743     above
1744     .\"
1745 nigel 93 for further details of the handling of digits following a backslash. There is
1746     no such problem when named parentheses are used. A back reference to any
1747     subpattern is possible using named parentheses (see below).
1748 nigel 75 .P
1749 nigel 93 Another way of avoiding the ambiguity inherent in the use of digits following a
1750 ph10 572 backslash is to use the \eg escape sequence. This escape must be followed by an
1751     unsigned number or a negative number, optionally enclosed in braces. These
1752     examples are all identical:
1753 nigel 93 .sp
1754     (ring), \e1
1755     (ring), \eg1
1756     (ring), \eg{1}
1757     .sp
1758 ph10 208 An unsigned number specifies an absolute reference without the ambiguity that
1759     is present in the older syntax. It is also useful when literal digits follow
1760     the reference. A negative number is a relative reference. Consider this
1761     example:
1762 nigel 93 .sp
1763     (abc(def)ghi)\eg{-1}
1764     .sp
1765     The sequence \eg{-1} is a reference to the most recently started capturing
1766 ph10 572 subpattern before \eg, that is, is it equivalent to \e2 in this example.
1767     Similarly, \eg{-2} would be equivalent to \e1. The use of relative references
1768     can be helpful in long patterns, and also in patterns that are created by
1769     joining together fragments that contain references within themselves.
1770 nigel 93 .P
1771 nigel 63 A back reference matches whatever actually matched the capturing subpattern in
1772     the current subject string, rather than anything matching the subpattern
1773     itself (see
1774     .\" HTML <a href="#subpatternsassubroutines">
1775     .\" </a>
1776     "Subpatterns as subroutines"
1777     .\"
1778     below for a way of doing that). So the pattern
1779 nigel 75 .sp
1780     (sens|respons)e and \e1ibility
1781     .sp
1782 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1783     "sense and responsibility". If caseful matching is in force at the time of the
1784     back reference, the case of letters is relevant. For example,
1785 nigel 75 .sp
1786     ((?i)rah)\es+\e1
1787     .sp
1788 nigel 63 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1789     capturing subpattern is matched caselessly.
1790 nigel 75 .P
1791 ph10 171 There are several different ways of writing back references to named
1792     subpatterns. The .NET syntax \ek{name} and the Perl syntax \ek<name> or
1793     \ek'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1794     back reference syntax, in which \eg can be used for both numeric and named
1795     references, is also supported. We could rewrite the above example in any of
1796 nigel 93 the following ways:
1797 nigel 75 .sp
1798 nigel 93 (?<p1>(?i)rah)\es+\ek<p1>
1799 ph10 171 (?'p1'(?i)rah)\es+\ek{p1}
1800 nigel 91 (?P<p1>(?i)rah)\es+(?P=p1)
1801 ph10 171 (?<p1>(?i)rah)\es+\eg{p1}
1802 nigel 75 .sp
1803 nigel 91 A subpattern that is referenced by name may appear in the pattern before or
1804     after the reference.
1805     .P
1806 nigel 63 There may be more than one back reference to the same subpattern. If a
1807     subpattern has not actually been used in a particular match, any back
1808 ph10 456 references to it always fail by default. For example, the pattern
1809 nigel 75 .sp
1810     (a|(bc))\e2
1811     .sp
1812 ph10 461 always fails if it starts to match "a" rather than "bc". However, if the
1813     PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back reference to an
1814 ph10 456 unset value matches an empty string.
1815     .P
1816     Because there may be many capturing parentheses in a pattern, all digits
1817     following a backslash are taken as part of a potential back reference number.
1818     If the pattern continues with a digit character, some delimiter must be used to
1819     terminate the back reference. If the PCRE_EXTENDED option is set, this can be
1820     whitespace. Otherwise, the \eg{ syntax or an empty comment (see
1821 nigel 75 .\" HTML <a href="#comments">
1822     .\" </a>
1823     "Comments"
1824     .\"
1825     below) can be used.
1826 ph10 488 .
1827     .SS "Recursive back references"
1828     .rs
1829     .sp
1830 nigel 63 A back reference that occurs inside the parentheses to which it refers fails
1831 nigel 75 when the subpattern is first used, so, for example, (a\e1) never matches.
1832 nigel 63 However, such references can be useful inside repeated subpatterns. For
1833     example, the pattern
1834 nigel 75 .sp
1835     (a|b\e1)+
1836     .sp
1837 nigel 63 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1838     the subpattern, the back reference matches the character string corresponding
1839     to the previous iteration. In order for this to work, the pattern must be such
1840     that the first iteration does not need to match the back reference. This can be
1841     done using alternation, as in the example above, or by a quantifier with a
1842     minimum of zero.
1843 ph10 488 .P
1844     Back references of this type cause the group that they reference to be treated
1845     as an
1846     .\" HTML <a href="#atomicgroup">
1847     .\" </a>
1848     atomic group.
1849     .\"
1850     Once the whole group has been matched, a subsequent matching failure cannot
1851     cause backtracking into the middle of the group.
1852 nigel 75 .
1853     .
1854     .\" HTML <a name="bigassertions"></a>
1855 nigel 63 .SH ASSERTIONS
1856     .rs
1857     .sp
1858     An assertion is a test on the characters following or preceding the current
1859     matching point that does not actually consume any characters. The simple
1860 nigel 75 assertions coded as \eb, \eB, \eA, \eG, \eZ, \ez, ^ and $ are described
1861     .\" HTML <a href="#smallassertions">
1862     .\" </a>
1863     above.
1864     .\"
1865     .P
1866 nigel 63 More complicated assertions are coded as subpatterns. There are two kinds:
1867     those that look ahead of the current position in the subject string, and those
1868 nigel 75 that look behind it. An assertion subpattern is matched in the normal way,
1869     except that it does not cause the current matching position to be changed.
1870     .P
1871 ph10 637 Assertion subpatterns are not capturing subpatterns. If such an assertion
1872     contains capturing subpatterns within it, these are counted for the purposes of
1873     numbering the capturing subpatterns in the whole pattern. However, substring
1874     capturing is carried out only for positive assertions, because it does not make
1875     sense for negative assertions.
1876     .P
1877 ph10 643 For compatibility with Perl, assertion subpatterns may be repeated; though
1878 ph10 654 it makes no sense to assert the same thing several times, the side effect of
1879 ph10 643 capturing parentheses may occasionally be useful. In practice, there only three
1880     cases:
1881 ph10 637 .sp
1882 ph10 654 (1) If the quantifier is {0}, the assertion is never obeyed during matching.
1883     However, it may contain internal capturing parenthesized groups that are called
1884 ph10 637 from elsewhere via the
1885     .\" HTML <a href="#subpatternsassubroutines">
1886     .\" </a>
1887     subroutine mechanism.
1888     .\"
1889     .sp
1890 ph10 654 (2) If quantifier is {0,n} where n is greater than zero, it is treated as if it
1891     were {0,1}. At run time, the rest of the pattern match is tried with and
1892 ph10 637 without the assertion, the order depending on the greediness of the quantifier.
1893     .sp
1894 ph10 654 (3) If the minimum repetition is greater than zero, the quantifier is ignored.
1895 ph10 637 The assertion is obeyed just once when encountered during matching.
1896 nigel 75 .
1897     .
1898     .SS "Lookahead assertions"
1899     .rs
1900     .sp
1901 nigel 91 Lookahead assertions start with (?= for positive assertions and (?! for
1902     negative assertions. For example,
1903 nigel 75 .sp
1904     \ew+(?=;)
1905     .sp
1906 nigel 63 matches a word followed by a semicolon, but does not include the semicolon in
1907     the match, and
1908 nigel 75 .sp
1909 nigel 63 foo(?!bar)
1910 nigel 75 .sp
1911 nigel 63 matches any occurrence of "foo" that is not followed by "bar". Note that the
1912     apparently similar pattern
1913 nigel 75 .sp
1914 nigel 63 (?!foo)bar
1915 nigel 75 .sp
1916 nigel 63 does not find an occurrence of "bar" that is preceded by something other than
1917     "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1918     (?!foo) is always true when the next three characters are "bar". A
1919 nigel 75 lookbehind assertion is needed to achieve the other effect.
1920     .P
1921 nigel 63 If you want to force a matching failure at some point in a pattern, the most
1922     convenient way to do it is with (?!) because an empty string always matches, so
1923     an assertion that requires there not to be an empty string must always fail.
1924 ph10 572 The backtracking control verb (*FAIL) or (*F) is a synonym for (?!).
1925 nigel 75 .
1926     .
1927     .\" HTML <a name="lookbehind"></a>
1928     .SS "Lookbehind assertions"
1929     .rs
1930     .sp
1931 nigel 63 Lookbehind assertions start with (?<= for positive assertions and (?<! for
1932     negative assertions. For example,
1933 nigel 75 .sp
1934 nigel 63 (?<!foo)bar
1935 nigel 75 .sp
1936 nigel 63 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1937     a lookbehind assertion are restricted such that all the strings it matches must
1938 nigel 91 have a fixed length. However, if there are several top-level alternatives, they
1939     do not all have to have the same fixed length. Thus
1940 nigel 75 .sp
1941 nigel 63 (?<=bullock|donkey)
1942 nigel 75 .sp
1943 nigel 63 is permitted, but
1944 nigel 75 .sp
1945 nigel 63 (?<!dogs?|cats?)
1946 nigel 75 .sp
1947 nigel 63 causes an error at compile time. Branches that match different length strings
1948     are permitted only at the top level of a lookbehind assertion. This is an
1949 ph10 572 extension compared with Perl, which requires all branches to match the same
1950     length of string. An assertion such as
1951 nigel 75 .sp
1952 nigel 63 (?<=ab(c|de))
1953 nigel 75 .sp
1954 nigel 63 is not permitted, because its single top-level branch can match two different
1955 ph10 454 lengths, but it is acceptable to PCRE if rewritten to use two top-level
1956     branches:
1957 nigel 75 .sp
1958 nigel 63 (?<=abc|abde)
1959 nigel 75 .sp
1960 ph10 572 In some cases, the escape sequence \eK
1961 ph10 168 .\" HTML <a href="#resetmatchstart">
1962     .\" </a>
1963     (see above)
1964     .\"
1965 ph10 461 can be used instead of a lookbehind assertion to get round the fixed-length
1966 ph10 454 restriction.
1967 ph10 168 .P
1968 nigel 63 The implementation of lookbehind assertions is, for each alternative, to
1969 nigel 93 temporarily move the current position back by the fixed length and then try to
1970 nigel 63 match. If there are insufficient characters before the current position, the
1971 nigel 93 assertion fails.
1972 nigel 75 .P
1973     PCRE does not allow the \eC escape (which matches a single byte in UTF-8 mode)
1974 nigel 63 to appear in lookbehind assertions, because it makes it impossible to calculate
1975 nigel 93 the length of the lookbehind. The \eX and \eR escapes, which can match
1976     different numbers of bytes, are also not permitted.
1977 nigel 75 .P
1978 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
1979     .\" </a>
1980     "Subroutine"
1981     .\"
1982     calls (see below) such as (?2) or (?&X) are permitted in lookbehinds, as long
1983 ph10 461 as the subpattern matches a fixed-length string.
1984 ph10 454 .\" HTML <a href="#recursion">
1985     .\" </a>
1986     Recursion,
1987     .\"
1988     however, is not supported.
1989     .P
1990 nigel 93 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1991 ph10 456 specify efficient matching of fixed-length strings at the end of subject
1992     strings. Consider a simple pattern such as
1993 nigel 75 .sp
1994 nigel 63 abcd$
1995 nigel 75 .sp
1996 nigel 63 when applied to a long string that does not match. Because matching proceeds
1997     from left to right, PCRE will look for each "a" in the subject and then see if
1998     what follows matches the rest of the pattern. If the pattern is specified as
1999 nigel 75 .sp
2000 nigel 63 ^.*abcd$
2001 nigel 75 .sp
2002 nigel 63 the initial .* matches the entire string at first, but when this fails (because
2003     there is no following "a"), it backtracks to match all but the last character,
2004     then all but the last two characters, and so on. Once again the search for "a"
2005     covers the entire string, from right to left, so we are no better off. However,
2006     if the pattern is written as
2007 nigel 75 .sp
2008 nigel 63 ^.*+(?<=abcd)
2009 nigel 75 .sp
2010 nigel 93 there can be no backtracking for the .*+ item; it can match only the entire
2011 nigel 63 string. The subsequent lookbehind assertion does a single test on the last four
2012     characters. If it fails, the match fails immediately. For long strings, this
2013     approach makes a significant difference to the processing time.
2014 nigel 75 .
2015     .
2016     .SS "Using multiple assertions"
2017     .rs
2018     .sp
2019 nigel 63 Several assertions (of any sort) may occur in succession. For example,
2020 nigel 75 .sp
2021     (?<=\ed{3})(?<!999)foo
2022     .sp
2023 nigel 63 matches "foo" preceded by three digits that are not "999". Notice that each of
2024     the assertions is applied independently at the same point in the subject
2025     string. First there is a check that the previous three characters are all
2026     digits, and then there is a check that the same three characters are not "999".
2027 nigel 75 This pattern does \fInot\fP match "foo" preceded by six characters, the first
2028 nigel 63 of which are digits and the last three of which are not "999". For example, it
2029     doesn't match "123abcfoo". A pattern to do that is
2030 nigel 75 .sp
2031     (?<=\ed{3}...)(?<!999)foo
2032     .sp
2033 nigel 63 This time the first assertion looks at the preceding six characters, checking
2034     that the first three are digits, and then the second assertion checks that the
2035     preceding three characters are not "999".
2036 nigel 75 .P
2037 nigel 63 Assertions can be nested in any combination. For example,
2038 nigel 75 .sp
2039 nigel 63 (?<=(?<!foo)bar)baz
2040 nigel 75 .sp
2041 nigel 63 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
2042     preceded by "foo", while
2043 nigel 75 .sp
2044     (?<=\ed{3}(?!999)...)foo
2045     .sp
2046     is another pattern that matches "foo" preceded by three digits and any three
2047 nigel 63 characters that are not "999".
2048 nigel 75 .
2049     .
2050 nigel 91 .\" HTML <a name="conditions"></a>
2052 nigel 63 .rs
2053     .sp
2054     It is possible to cause the matching process to obey a subpattern
2055     conditionally or to choose between two alternative subpatterns, depending on
2056 ph10 461 the result of an assertion, or whether a specific capturing subpattern has
2057 ph10 456 already been matched. The two possible forms of conditional subpattern are:
2058 nigel 75 .sp
2059 nigel 63 (?(condition)yes-pattern)
2060     (?(condition)yes-pattern|no-pattern)
2061 nigel 75 .sp
2062 nigel 63 If the condition is satisfied, the yes-pattern is used; otherwise the
2063     no-pattern (if present) is used. If there are more than two alternatives in the
2064 ph10 557 subpattern, a compile-time error occurs. Each of the two alternatives may
2065 ph10 579 itself contain nested subpatterns of any form, including conditional
2066 ph10 557 subpatterns; the restriction to two alternatives applies only at the level of
2067 ph10 579 the condition. This pattern fragment is an example where the alternatives are
2068 ph10 557 complex:
2069     .sp
2070     (?(1) (A|B|C) | (D | (?(2)E|F) | E) )
2071     .sp
2072 nigel 75 .P
2073 nigel 93 There are four kinds of condition: references to subpatterns, references to
2074     recursion, a pseudo-condition called DEFINE, and assertions.
2075     .
2076     .SS "Checking for a used subpattern by number"
2077     .rs
2078     .sp
2079     If the text between the parentheses consists of a sequence of digits, the
2080 ph10 456 condition is true if a capturing subpattern of that number has previously
2081 ph10 461 matched. If there is more than one capturing subpattern with the same number
2082     (see the earlier
2083 ph10 456 .\"
2084     .\" HTML <a href="#recursion">
2085     .\" </a>
2086     section about duplicate subpattern numbers),
2087     .\"
2088 ph10 572 the condition is true if any of them have matched. An alternative notation is
2089 ph10 456 to precede the digits with a plus or minus sign. In this case, the subpattern
2090     number is relative rather than absolute. The most recently opened parentheses
2091 ph10 572 can be referenced by (?(-1), the next most recent by (?(-2), and so on. Inside
2092 ph10 579 loops it can also make sense to refer to subsequent groups. The next
2093 ph10 572 parentheses to be opened can be referenced as (?(+1), and so on. (The value
2094     zero in any of these forms is not used; it provokes a compile-time error.)
2095 nigel 91 .P
2096     Consider the following pattern, which contains non-significant white space to
2097     make it more readable (assume the PCRE_EXTENDED option) and to divide it into
2098     three parts for ease of discussion:
2099 nigel 75 .sp
2100     ( \e( )? [^()]+ (?(1) \e) )
2101     .sp
2102 nigel 63 The first part matches an optional opening parenthesis, and if that
2103     character is present, sets it as the first captured substring. The second part
2104     matches one or more characters that are not parentheses. The third part is a
2105 ph10 572 conditional subpattern that tests whether or not the first set of parentheses
2106     matched. If they did, that is, if subject started with an opening parenthesis,
2107 nigel 63 the condition is true, and so the yes-pattern is executed and a closing
2108     parenthesis is required. Otherwise, since no-pattern is not present, the
2109     subpattern matches nothing. In other words, this pattern matches a sequence of
2110 nigel 93 non-parentheses, optionally enclosed in parentheses.
2111 ph10 167 .P
2112 ph10 172 If you were embedding this pattern in a larger one, you could use a relative
2113 ph10 167 reference:
2114     .sp
2115     ...other stuff... ( \e( )? [^()]+ (?(-1) \e) ) ...
2116     .sp
2117     This makes the fragment independent of the parentheses in the larger pattern.
2118 nigel 93 .
2119     .SS "Checking for a used subpattern by name"
2120     .rs
2121 nigel 91 .sp
2122 nigel 93 Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a used
2123     subpattern by name. For compatibility with earlier versions of PCRE, which had
2124     this facility before Perl, the syntax (?(name)...) is also recognized. However,
2125     there is a possible ambiguity with this syntax, because subpattern names may
2126     consist entirely of digits. PCRE looks first for a named subpattern; if it
2127     cannot find one and the name consists entirely of digits, PCRE looks for a
2128     subpattern of that number, which must be greater than zero. Using subpattern
2129     names that consist entirely of digits is not recommended.
2130     .P
2131     Rewriting the above example to use a named subpattern gives this:
2132 nigel 91 .sp
2133 nigel 93 (?<OPEN> \e( )? [^()]+ (?(<OPEN>) \e) )
2134     .sp
2135 ph10 461 If the name used in a condition of this kind is a duplicate, the test is
2136     applied to all subpatterns of the same name, and is true if any one of them has
2137 ph10 459 matched.
2138 nigel 93 .
2139     .SS "Checking for pattern recursion"
2140     .rs
2141     .sp
2142 nigel 91 If the condition is the string (R), and there is no subpattern with the name R,
2143 nigel 93 the condition is true if a recursive call to the whole pattern or any
2144     subpattern has been made. If digits or a name preceded by ampersand follow the
2145     letter R, for example:
2146     .sp
2147     (?(R3)...) or (?(R&name)...)
2148     .sp
2149 ph10 456 the condition is true if the most recent recursion is into a subpattern whose
2150 nigel 93 number or name is given. This condition does not check the entire recursion
2151 ph10 461 stack. If the name used in a condition of this kind is a duplicate, the test is
2152     applied to all subpatterns of the same name, and is true if any one of them is
2153     the most recent recursion.
2154 nigel 75 .P
2155 ph10 461 At "top level", all these recursion test conditions are false.
2156 ph10 454 .\" HTML <a href="#recursion">
2157     .\" </a>
2158 ph10 459 The syntax for recursive patterns
2159 ph10 454 .\"
2160 ph10 459 is described below.
2161 nigel 93 .
2162 ph10 572 .\" HTML <a name="subdefine"></a>
2163 nigel 93 .SS "Defining subpatterns for use by reference only"
2164     .rs
2165     .sp
2166     If the condition is the string (DEFINE), and there is no subpattern with the
2167     name DEFINE, the condition is always false. In this case, there may be only one
2168     alternative in the subpattern. It is always skipped if control reaches this
2169     point in the pattern; the idea of DEFINE is that it can be used to define
2170 ph10 716 subroutines that can be referenced from elsewhere. (The use of
2171 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
2172     .\" </a>
2173 ph10 716 subroutines
2174 ph10 454 .\"
2175 ph10 572 is described below.) For example, a pattern to match an IPv4 address such as
2176     "" could be written like this (ignore whitespace and line
2177     breaks):
2178 nigel 93 .sp
2179     (?(DEFINE) (?<byte> 2[0-4]\ed | 25[0-5] | 1\ed\ed | [1-9]?\ed) )
2180     \eb (?&byte) (\e.(?&byte)){3} \eb
2181     .sp
2182     The first part of the pattern is a DEFINE group inside which a another group
2183     named "byte" is defined. This matches an individual component of an IPv4
2184     address (a number less than 256). When matching takes place, this part of the
2185 ph10 456 pattern is skipped because DEFINE acts like a false condition. The rest of the
2186     pattern uses references to the named group to match the four dot-separated
2187     components of an IPv4 address, insisting on a word boundary at each end.
2188 nigel 93 .
2189     .SS "Assertion conditions"
2190     .rs
2191     .sp
2192     If the condition is not in any of the above formats, it must be an assertion.
2193 nigel 63 This may be a positive or negative lookahead or lookbehind assertion. Consider
2194     this pattern, again containing non-significant white space, and with the two
2195     alternatives on the second line:
2196 nigel 75 .sp
2197 nigel 63 (?(?=[^a-z]*[a-z])
2198 nigel 75 \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
2199     .sp
2200 nigel 63 The condition is a positive lookahead assertion that matches an optional
2201     sequence of non-letters followed by a letter. In other words, it tests for the
2202     presence of at least one letter in the subject. If a letter is found, the
2203     subject is matched against the first alternative; otherwise it is matched
2204     against the second. This pattern matches strings in one of the two forms
2205     dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
2206 nigel 75 .
2207     .
2208     .\" HTML <a name="comments"></a>
2209 nigel 63 .SH COMMENTS
2210     .rs
2211     .sp
2212 ph10 579 There are two ways of including comments in patterns that are processed by
2213 ph10 562 PCRE. In both cases, the start of the comment must not be in a character class,
2214     nor in the middle of any other sequence of related characters such as (?: or a
2215     subpattern name or number. The characters that make up a comment play no part
2216     in the pattern matching.
2217     .P
2218 nigel 75 The sequence (?# marks the start of a comment that continues up to the next
2219 ph10 562 closing parenthesis. Nested parentheses are not permitted. If the PCRE_EXTENDED
2220     option is set, an unescaped # character also introduces a comment, which in
2221     this case continues to immediately after the next newline character or
2222     character sequence in the pattern. Which characters are interpreted as newlines
2223     is controlled by the options passed to \fBpcre_compile()\fP or by a special
2224     sequence at the start of the pattern, as described in the section entitled
2225 ph10 572 .\" HTML <a href="#newlines">
2226 ph10 556 .\" </a>
2227     "Newline conventions"
2228     .\"
2229 ph10 572 above. Note that the end of this type of comment is a literal newline sequence
2230     in the pattern; escape sequences that happen to represent a newline do not
2231     count. For example, consider this pattern when PCRE_EXTENDED is set, and the
2232     default newline convention is in force:
2233 ph10 556 .sp
2234     abc #comment \en still comment
2235     .sp
2236 ph10 579 On encountering the # character, \fBpcre_compile()\fP skips along, looking for
2237 ph10 556 a newline in the pattern. The sequence \en is still literal at this stage, so
2238     it does not terminate the comment. Only an actual character with the code value
2239 ph10 562 0x0a (the default newline) does so.
2240 nigel 75 .
2241     .
2242 nigel 91 .\" HTML <a name="recursion"></a>
2243 nigel 75 .SH "RECURSIVE PATTERNS"
2244 nigel 63 .rs
2245     .sp
2246     Consider the problem of matching a string in parentheses, allowing for
2247     unlimited nested parentheses. Without the use of recursion, the best that can
2248     be done is to use a pattern that matches up to some fixed depth of nesting. It
2249 nigel 93 is not possible to handle an arbitrary nesting depth.
2250     .P
2251     For some time, Perl has provided a facility that allows regular expressions to
2252     recurse (amongst other things). It does this by interpolating Perl code in the
2253     expression at run time, and the code can refer to the expression itself. A Perl
2254     pattern using code interpolation to solve the parentheses problem can be
2255     created like this:
2256 nigel 75 .sp
2257     $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
2258     .sp
2259 nigel 63 The (?p{...}) item interpolates Perl code at run time, and in this case refers
2260 nigel 93 recursively to the pattern in which it appears.
2261 nigel 75 .P
2262 nigel 93 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
2263     supports special syntax for recursion of the entire pattern, and also for
2264     individual subpattern recursion. After its introduction in PCRE and Python,
2265 ph10 453 this kind of recursion was subsequently introduced into Perl at release 5.10.
2266 nigel 75 .P
2267 nigel 93 A special item that consists of (? followed by a number greater than zero and a
2268 ph10 716 closing parenthesis is a recursive subroutine call of the subpattern of the
2269     given number, provided that it occurs inside that subpattern. (If not, it is a
2270 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
2271     .\" </a>
2272 ph10 716 non-recursive subroutine
2273 ph10 454 .\"
2274 nigel 93 call, which is described in the next section.) The special item (?R) or (?0) is
2275     a recursive call of the entire regular expression.
2276 nigel 87 .P
2277     This PCRE pattern solves the nested parentheses problem (assume the
2278     PCRE_EXTENDED option is set so that white space is ignored):
2279 nigel 75 .sp
2280 ph10 456 \e( ( [^()]++ | (?R) )* \e)
2281 nigel 75 .sp
2282 nigel 63 First it matches an opening parenthesis. Then it matches any number of
2283     substrings which can either be a sequence of non-parentheses, or a recursive
2284 nigel 87 match of the pattern itself (that is, a correctly parenthesized substring).
2285 ph10 461 Finally there is a closing parenthesis. Note the use of a possessive quantifier
2286 ph10 456 to avoid backtracking into sequences of non-parentheses.
2287 nigel 75 .P
2288 nigel 63 If this were part of a larger pattern, you would not want to recurse the entire
2289     pattern, so instead you could use this:
2290 nigel 75 .sp
2291 ph10 456 ( \e( ( [^()]++ | (?1) )* \e) )
2292 nigel 75 .sp
2293 nigel 63 We have put the pattern into parentheses, and caused the recursion to refer to
2294 ph10 172 them instead of the whole pattern.
2295 ph10 166 .P
2296     In a larger pattern, keeping track of parenthesis numbers can be tricky. This
2297 ph10 572 is made easier by the use of relative references. Instead of (?1) in the
2298     pattern above you can write (?-2) to refer to the second most recently opened
2299     parentheses preceding the recursion. In other words, a negative number counts
2300     capturing parentheses leftwards from the point at which it is encountered.
2301 ph10 166 .P
2302     It is also possible to refer to subsequently opened parentheses, by writing
2303     references such as (?+2). However, these cannot be recursive because the
2304     reference is not inside the parentheses that are referenced. They are always
2305 ph10 454 .\" HTML <a href="#subpatternsassubroutines">
2306     .\" </a>
2307 ph10 716 non-recursive subroutine
2308 ph10 454 .\"
2309     calls, as described in the next section.
2310 ph10 166 .P
2311     An alternative approach is to use named parentheses instead. The Perl syntax
2312     for this is (?&name); PCRE's earlier syntax (?P>name) is also supported. We
2313     could rewrite the above example as follows:
2314 nigel 75 .sp
2315 ph10 456 (?<pn> \e( ( [^()]++ | (?&pn) )* \e) )
2316 nigel 75 .sp
2317 nigel 93 If there is more than one subpattern with the same name, the earliest one is
2318 ph10 172 used.
2319 ph10 166 .P
2320     This particular example pattern that we have been looking at contains nested
2321 ph10 456 unlimited repeats, and so the use of a possessive quantifier for matching
2322     strings of non-parentheses is important when applying the pattern to strings
2323     that do not match. For example, when this pattern is applied to
2324 nigel 75 .sp
2325 nigel 63 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
2326 nigel 75 .sp
2327 ph10 456 it yields "no match" quickly. However, if a possessive quantifier is not used,
2328 nigel 63 the match runs for a very long time indeed because there are so many different
2329     ways the + and * repeats can carve up the subject, and all have to be tested
2330     before failure can be reported.
2331 nigel 75 .P
2332 ph10 464 At the end of a match, the values of capturing parentheses are those from
2333     the outermost level. If you want to obtain intermediate values, a callout
2334     function can be used (see below and the
2335 nigel 63 .\" HREF
2336 nigel 75 \fBpcrecallout\fP
2337 nigel 63 .\"
2338     documentation). If the pattern above is matched against
2339 nigel 75 .sp
2340 nigel 63 (ab(cd)ef)
2341 nigel 75 .sp
2342 ph10 464 the value for the inner capturing parentheses (numbered 2) is "ef", which is
2343     the last value taken on at the top level. If a capturing subpattern is not
2344 ph10 724 matched at the top level, its final captured value is unset, even if it was
2345     (temporarily) set at a deeper level during the matching process.
2346 nigel 75 .P
2347 ph10 464 If there are more than 15 capturing parentheses in a pattern, PCRE has to
2348     obtain extra memory to store data during a recursion, which it does by using
2349     \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no memory can
2350     be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
2351     .P
2352 nigel 63 Do not confuse the (?R) item with the condition (R), which tests for recursion.
2353     Consider this pattern, which matches text in angle brackets, allowing for
2354     arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2355     recursing), whereas any characters are permitted at the outer level.
2356 nigel 75 .sp
2357     < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
2358     .sp
2359 nigel 63 In this pattern, (?(R) is the start of a conditional subpattern, with two
2360     different alternatives for the recursive and non-recursive cases. The (?R) item
2361     is the actual recursive call.
2362 nigel 75 .
2363     .
2364 ph10 453 .\" HTML <a name="recursiondifference"></a>
2365 ph10 724 .SS "Differences in recursion processing between PCRE and Perl"
2366 ph10 453 .rs
2367     .sp
2368 ph10 724 Recursion processing in PCRE differs from Perl in two important ways. In PCRE
2369     (like Python, but unlike Perl), a recursive subpattern call is always treated
2370     as an atomic group. That is, once it has matched some of the subject string, it
2371     is never re-entered, even if it contains untried alternatives and there is a
2372     subsequent matching failure. This can be illustrated by the following pattern,
2373     which purports to match a palindromic string that contains an odd number of
2374     characters (for example, "a", "aba", "abcba", "abcdcba"):
2375 ph10 453 .sp
2376     ^(.|(.)(?1)\e2)$
2377     .sp
2378 ph10 461 The idea is that it either matches a single character, or two identical
2379     characters surrounding a sub-palindrome. In Perl, this pattern works; in PCRE
2380 ph10 453 it does not if the pattern is longer than three characters. Consider the
2381     subject string "abcba":
2382     .P
2383 ph10 461 At the top level, the first character is matched, but as it is not at the end
2384 ph10 453 of the string, the first alternative fails; the second alternative is taken
2385     and the recursion kicks in. The recursive call to subpattern 1 successfully
2386     matches the next character ("b"). (Note that the beginning and end of line
2387     tests are not part of the recursion).
2388     .P
2389     Back at the top level, the next character ("c") is compared with what
2390 ph10 461 subpattern 2 matched, which was "a". This fails. Because the recursion is
2391 ph10 453 treated as an atomic group, there are now no backtracking points, and so the
2392     entire match fails. (Perl is able, at this point, to re-enter the recursion and
2393     try the second alternative.) However, if the pattern is written with the
2394     alternatives in the other order, things are different:
2395     .sp
2396     ^((.)(?1)\e2|.)$
2397     .sp
2398 ph10 461 This time, the recursing alternative is tried first, and continues to recurse
2399     until it runs out of characters, at which point the recursion fails. But this
2400     time we do have another alternative to try at the higher level. That is the big
2401 ph10 453 difference: in the previous case the remaining alternative is at a deeper
2402     recursion level, which PCRE cannot use.
2403     .P
2404 ph10 572 To change the pattern so that it matches all palindromic strings, not just
2405     those with an odd number of characters, it is tempting to change the pattern to
2406     this:
2407 ph10 453 .sp
2408     ^((.)(?1)\e2|.?)$
2409     .sp
2410 ph10 461 Again, this works in Perl, but not in PCRE, and for the same reason. When a
2411     deeper recursion has matched a single character, it cannot be entered again in
2412     order to match an empty string. The solution is to separate the two cases, and
2413 ph10 453 write out the odd and even cases as alternatives at the higher level:
2414     .sp
2415     ^(?:((.)(?1)\e2|)|((.)(?3)\e4|.))
2416 ph10 461 .sp
2417     If you want to match typical palindromic phrases, the pattern has to ignore all
2418 ph10 453 non-word characters, which can be done like this:
2419     .sp
2420 ph10 461 ^\eW*+(?:((.)\eW*+(?1)\eW*+\e2|)|((.)\eW*+(?3)\eW*+\e4|\eW*+.\eW*+))\eW*+$
2421 ph10 453 .sp
2422 ph10 461 If run with the PCRE_CASELESS option, this pattern matches phrases such as "A
2423     man, a plan, a canal: Panama!" and it works well in both PCRE and Perl. Note
2424     the use of the possessive quantifier *+ to avoid backtracking into sequences of
2425 ph10 453 non-word characters. Without this, PCRE takes a great deal longer (ten times or
2426     more) to match typical phrases, and Perl takes so long that you think it has
2427     gone into a loop.
2428 ph10 456 .P
2429     \fBWARNING\fP: The palindrome-matching patterns above work only if the subject
2430     string does not start with a palindrome that is shorter than the entire string.
2431     For example, although "abcba" is correctly matched, if the subject is "ababa",
2432     PCRE finds the palindrome "aba" at the start, then fails at top level because
2433     the end of the string does not follow. Once again, it cannot jump back into the
2434     recursion to try other alternatives, so the entire match fails.
2435 ph10 724 .P
2436 ph10 733 The second way in which PCRE and Perl differ in their recursion processing is
2437     in the handling of captured values. In Perl, when a subpattern is called
2438     recursively or as a subpattern (see the next section), it has no access to any
2439     values that were captured outside the recursion, whereas in PCRE these values
2440 ph10 724 can be referenced. Consider this pattern:
2441     .sp
2442     ^(.)(\e1|a(?2))
2443     .sp
2444 ph10 733 In PCRE, this pattern matches "bab". The first capturing parentheses match "b",
2445 ph10 724 then in the second group, when the back reference \e1 fails to match "b", the
2446     second alternative matches "a" and then recurses. In the recursion, \e1 does
2447     now match "b" and so the whole match succeeds. In Perl, the pattern fails to
2448     match because inside the recursive call \e1 cannot access the externally set
2449     value.
2450 ph10 453 .
2451     .
2452 nigel 63 .\" HTML <a name="subpatternsassubroutines"></a>
2454 nigel 63 .rs
2455     .sp
2456 ph10 716 If the syntax for a recursive subpattern call (either by number or by
2457 nigel 63 name) is used outside the parentheses to which it refers, it operates like a
2458 ph10 716 subroutine in a programming language. The called subpattern may be defined
2459 ph10 166 before or after the reference. A numbered reference can be absolute or
2460     relative, as in these examples:
2461 nigel 75 .sp
2462 ph10 166 (...(absolute)...)...(?2)...
2463     (...(relative)...)...(?-1)...
2464 ph10 172 (...(?+1)...(relative)...
2465 ph10 166 .sp
2466     An earlier example pointed out that the pattern
2467     .sp
2468 nigel 75 (sens|respons)e and \e1ibility
2469     .sp
2470 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
2471     "sense and responsibility". If instead the pattern
2472 nigel 75 .sp
2473 nigel 63 (sens|respons)e and (?1)ibility
2474 nigel 75 .sp
2475 nigel 63 is used, it does match "sense and responsibility" as well as the other two
2476 nigel 93 strings. Another example is given in the discussion of DEFINE above.
2477 nigel 87 .P
2478 ph10 716 All subroutine calls, whether recursive or not, are always treated as atomic
2479     groups. That is, once a subroutine has matched some of the subject string, it
2480     is never re-entered, even if it contains untried alternatives and there is a
2481     subsequent matching failure. Any capturing parentheses that are set during the
2482     subroutine call revert to their previous values afterwards.
2483 nigel 93 .P
2484 ph10 716 Processing options such as case-independence are fixed when a subpattern is
2485     defined, so if it is used as a subroutine, such options cannot be changed for
2486     different calls. For example, consider this pattern:
2487 nigel 93 .sp
2488 ph10 166 (abc)(?i:(?-1))
2489 nigel 93 .sp
2490     It matches "abcabc". It does not match "abcABC" because the change of
2491     processing option does not affect the called subpattern.
2492 nigel 75 .
2493     .
2494 ph10 333 .\" HTML <a name="onigurumasubroutines"></a>
2496     .rs
2497     .sp
2498 ph10 345 For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
2499     a number enclosed either in angle brackets or single quotes, is an alternative
2500     syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2501 ph10 333 are two of the examples used above, rewritten using this syntax:
2502     .sp
2503     (?<pn> \e( ( (?>[^()]+) | \eg<pn> )* \e) )
2504     (sens|respons)e and \eg'1'ibility
2505     .sp
2506 ph10 345 PCRE supports an extension to Oniguruma: if a number is preceded by a
2507 ph10 333 plus or a minus sign it is taken as a relative reference. For example:
2508     .sp
2509     (abc)(?i:\eg<-1>)
2510     .sp
2511 ph10 345 Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
2512 ph10 333 synonymous. The former is a back reference; the latter is a subroutine call.
2513     .
2514     .
2515 nigel 63 .SH CALLOUTS
2516     .rs
2517     .sp
2518     Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2519     code to be obeyed in the middle of matching a regular expression. This makes it
2520     possible, amongst other things, to extract different substrings that match the
2521     same pair of parentheses when there is a repetition.
2522 nigel 75 .P
2523 nigel 63 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2524     code. The feature is called "callout". The caller of PCRE provides an external
2525 nigel 75 function by putting its entry point in the global variable \fIpcre_callout\fP.
2526 nigel 63 By default, this variable contains NULL, which disables all calling out.
2527 nigel 75 .P
2528 nigel 63 Within a regular expression, (?C) indicates the points at which the external
2529     function is to be called. If you want to identify different callout points, you
2530     can put a number less than 256 after the letter C. The default value is zero.
2531     For example, this pattern has two callout points:
2532 nigel 75 .sp
2533 ph10 155 (?C1)abc(?C2)def
2534 nigel 75 .sp
2535     If the PCRE_AUTO_CALLOUT flag is passed to \fBpcre_compile()\fP, callouts are
2536     automatically installed before each item in the pattern. They are all numbered
2537     255.
2538     .P
2539     During matching, when PCRE reaches a callout point (and \fIpcre_callout\fP is
2540 nigel 63 set), the external function is called. It is provided with the number of the
2541 nigel 75 callout, the position in the pattern, and, optionally, one item of data
2542     originally supplied by the caller of \fBpcre_exec()\fP. The callout function
2543     may cause matching to proceed, to backtrack, or to fail altogether. A complete
2544     description of the interface to the callout function is given in the
2545 nigel 63 .\" HREF
2546 nigel 75 \fBpcrecallout\fP
2547 nigel 63 .\"
2548     documentation.
2549 nigel 93 .
2550     .
2551 ph10 510 .\" HTML <a name="backtrackcontrol"></a>
2553 ph10 210 .rs
2554     .sp
2555 ph10 211 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2556 ph10 210 are described in the Perl documentation as "experimental and subject to change
2557 ph10 211 or removal in a future version of Perl". It goes on to say: "Their usage in
2558     production code should be noted to avoid problems during upgrades." The same
2559 ph10 210 remarks apply to the PCRE features described in this section.
2560     .P
2561 ph10 341 Since these verbs are specifically related to backtracking, most of them can be
2562     used only when the pattern is to be matched using \fBpcre_exec()\fP, which uses
2563 ph10 345 a backtracking algorithm. With the exception of (*FAIL), which behaves like a
2564 ph10 341 failing negative assertion, they cause an error if encountered by
2565 ph10 210 \fBpcre_dfa_exec()\fP.
2566     .P
2567 ph10 716 If any of these verbs are used in an assertion or in a subpattern that is
2568     called as a subroutine (whether or not recursively), their effect is confined
2569     to that subpattern; it does not extend to the surrounding pattern, with one
2570     exception: a *MARK that is encountered in a positive assertion \fIis\fP passed
2571     back (compare capturing parentheses in assertions). Note that such subpatterns
2572     are processed as anchored at the point where they are tested. Note also that
2573     Perl's treatment of subroutines is different in some cases.
2574 ph10 445 .P
2575 ph10 211 The new verbs make use of what was previously invalid syntax: an opening
2576 ph10 510 parenthesis followed by an asterisk. They are generally of the form
2577 ph10 512 (*VERB) or (*VERB:NAME). Some may take either form, with differing behaviour,
2578 ph10 716 depending on whether or not an argument is present. A name is any sequence of
2579     characters that does not include a closing parenthesis. If the name is empty,
2580     that is, if the closing parenthesis immediately follows the colon, the effect
2581     is as if the colon were not there. Any number of these verbs may occur in a
2582     pattern.
2583 ph10 510 .P
2584 ph10 512 PCRE contains some optimizations that are used to speed up matching by running
2585     some checks at the start of each match attempt. For example, it may know the
2586     minimum length of matching subject, or that a particular character must be
2587     present. When one of these optimizations suppresses the running of a match, any
2588     included backtracking verbs will not, of course, be processed. You can suppress
2589     the start-of-match optimizations by setting the PCRE_NO_START_OPTIMIZE option
2590 ph10 577 when calling \fBpcre_compile()\fP or \fBpcre_exec()\fP, or by starting the
2591     pattern with (*NO_START_OPT).
2592 ph10 210 .
2593 ph10 510 .
2594 ph10 210 .SS "Verbs that act immediately"
2595     .rs
2596     .sp
2597 ph10 512 The following verbs act as soon as they are encountered. They may not be
2598 ph10 510 followed by a name.
2599 ph10 210 .sp
2600     (*ACCEPT)
2601     .sp
2602     This verb causes the match to end successfully, skipping the remainder of the
2603 ph10 716 pattern. However, when it is inside a subpattern that is called as a
2604     subroutine, only that subpattern is ended successfully. Matching then continues
2605     at the outer level. If (*ACCEPT) is inside capturing parentheses, the data so
2606     far is captured. For example:
2607 ph10 210 .sp
2608 ph10 447 A((?:A|B(*ACCEPT)|C)D)
2609 ph10 210 .sp
2610 ph10 461 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is captured by
2611 ph10 447 the outer parentheses.
2612 ph10 210 .sp
2613     (*FAIL) or (*F)
2614     .sp
2615 ph10 716 This verb causes a matching failure, forcing backtracking to occur. It is
2616 ph10 210 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2617     probably useful only when combined with (?{}) or (??{}). Those are, of course,
2618     Perl features that are not present in PCRE. The nearest equivalent is the
2619     callout feature, as for example in this pattern:
2620     .sp
2621     a+(?C)(*FAIL)
2622     .sp
2623 ph10 211 A match with the string "aaaa" always fails, but the callout is taken before
2624     each backtrack happens (in this example, 10 times).
2625 ph10 210 .
2626 ph10 510 .
2627     .SS "Recording which path was taken"
2628     .rs
2629     .sp
2630 ph10 512 There is one verb whose main purpose is to track how a match was arrived at,
2631     though it also has a secondary use in conjunction with advancing the match
2632 ph10 510 starting point (see (*SKIP) below).
2633     .sp
2634     (*MARK:NAME) or (*:NAME)
2635     .sp
2636     A name is always required with this verb. There may be as many instances of
2637     (*MARK) as you like in a pattern, and their names do not have to be unique.
2638     .P
2639     When a match succeeds, the name of the last-encountered (*MARK) is passed back
2640     to the caller via the \fIpcre_extra\fP data structure, as described in the
2641     .\" HTML <a href="pcreapi.html#extradata">
2642     .\" </a>
2643     section on \fIpcre_extra\fP
2644     .\"
2645 ph10 512 in the
2646 ph10 510 .\" HREF
2647     \fBpcreapi\fP
2648     .\"
2649     documentation. No data is returned for a partial match. Here is an example of
2650     \fBpcretest\fP output, where the /K modifier requests the retrieval and
2651     outputting of (*MARK) data:
2652     .sp
2653     /X(*MARK:A)Y|X(*MARK:B)Z/K
2654     XY
2655     0: XY
2656     MK: A
2657     XZ
2658     0: XZ
2659     MK: B
2660     .sp
2661 ph10 512 The (*MARK) name is tagged with "MK:" in this output, and in this example it
2662     indicates which of the two alternatives matched. This is a more efficient way
2663 ph10 510 of obtaining this information than putting each alternative in its own
2664     capturing parentheses.
2665     .P
2666 ph10 654 If (*MARK) is encountered in a positive assertion, its name is recorded and
2667     passed back if it is the last-encountered. This does not happen for negative
2668 ph10 716 assertions.
2669 ph10 630 .P
2670 ph10 510 A name may also be returned after a failed match if the final path through the
2671     pattern involves (*MARK). However, unless (*MARK) used in conjunction with
2672     (*COMMIT), this is unlikely to happen for an unanchored pattern because, as the
2673     starting point for matching is advanced, the final check is often with an empty
2674     string, causing a failure before (*MARK) is reached. For example:
2675     .sp
2676     /X(*MARK:A)Y|X(*MARK:B)Z/K
2677     XP
2678     No match
2679     .sp
2680     There are three potential starting points for this match (starting with X,
2681 ph10 512 starting with P, and with an empty string). If the pattern is anchored, the
2682 ph10 510 result is different:
2683     .sp
2684     /^X(*MARK:A)Y|^X(*MARK:B)Z/K
2685     XP
2686     No match, mark = B
2687     .sp
2688 ph10 512 PCRE's start-of-match optimizations can also interfere with this. For example,
2689     if, as a result of a call to \fBpcre_study()\fP, it knows the minimum
2690 ph10 510 subject length for a match, a shorter subject will not be scanned at all.
2691     .P
2692 ph10 512 Note that similar anomalies (though different in detail) exist in Perl, no
2693     doubt for the same reasons. The use of (*MARK) data after a failed match of an
2694 ph10 510 unanchored pattern is not recommended, unless (*COMMIT) is involved.
2695     .
2696     .
2697 ph10 210 .SS "Verbs that act after backtracking"
2698     .rs
2699     .sp
2700 ph10 211 The following verbs do nothing when they are encountered. Matching continues
2701 ph10 510 with what follows, but if there is no subsequent match, causing a backtrack to
2702     the verb, a failure is forced. That is, backtracking cannot pass to the left of
2703     the verb. However, when one of these verbs appears inside an atomic group, its
2704     effect is confined to that group, because once the group has been matched,
2705     there is never any backtracking into it. In this situation, backtracking can
2706 ph10 512 "jump back" to the left of the entire atomic group. (Remember also, as stated
2707 ph10 510 above, that this localization also applies in subroutine calls and assertions.)
2708     .P
2709     These verbs differ in exactly what kind of failure occurs when backtracking
2710     reaches them.
2711 ph10 210 .sp
2712     (*COMMIT)
2713     .sp
2714 ph10 510 This verb, which may not be followed by a name, causes the whole match to fail
2715     outright if the rest of the pattern does not match. Even if the pattern is
2716     unanchored, no further attempts to find a match by advancing the starting point
2717     take place. Once (*COMMIT) has been passed, \fBpcre_exec()\fP is committed to
2718     finding a match at the current starting point, or not at all. For example:
2719 ph10 210 .sp
2720     a+(*COMMIT)b
2721     .sp
2722 ph10 211 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2723 ph10 512 dynamic anchor, or "I've started, so I must finish." The name of the most
2724     recently passed (*MARK) in the path is passed back when (*COMMIT) forces a
2725 ph10 510 match failure.
2726     .P
2727 ph10 512 Note that (*COMMIT) at the start of a pattern is not the same as an anchor,
2728     unless PCRE's start-of-match optimizations are turned off, as shown in this
2729 ph10 510 \fBpcretest\fP example:
2730 ph10 210 .sp
2731 ph10 510 /(*COMMIT)abc/
2732     xyzabc
2733     0: abc
2734     xyzabc\eY
2735     No match
2736 ph10 210 .sp
2737 ph10 512 PCRE knows that any match must start with "a", so the optimization skips along
2738 ph10 510 the subject to "a" before running the first match attempt, which succeeds. When
2739     the optimization is disabled by the \eY escape in the second subject, the match
2740 ph10 512 starts at "x" and so the (*COMMIT) causes it to fail without trying any other
2741 ph10 510 starting points.
2742 ph10 210 .sp
2743 ph10 510 (*PRUNE) or (*PRUNE:NAME)
2744     .sp
2745 ph10 512 This verb causes the match to fail at the current starting position in the
2746 ph10 510 subject if the rest of the pattern does not match. If the pattern is
2747     unanchored, the normal "bumpalong" advance to the next starting character then
2748     happens. Backtracking can occur as usual to the left of (*PRUNE), before it is
2749     reached, or when matching to the right of (*PRUNE), but if there is no match to
2750     the right, backtracking cannot cross (*PRUNE). In simple cases, the use of
2751     (*PRUNE) is just an alternative to an atomic group or possessive quantifier,
2752     but there are some uses of (*PRUNE) that cannot be expressed in any other way.
2753     The behaviour of (*PRUNE:NAME) is the same as (*MARK:NAME)(*PRUNE) when the
2754     match fails completely; the name is passed back if this is the final attempt.
2755     (*PRUNE:NAME) does not pass back a name if the match succeeds. In an anchored
2756     pattern (*PRUNE) has the same effect as (*COMMIT).
2757     .sp
2758 ph10 210 (*SKIP)
2759     .sp
2760 ph10 510 This verb, when given without a name, is like (*PRUNE), except that if the
2761     pattern is unanchored, the "bumpalong" advance is not to the next character,
2762     but to the position in the subject where (*SKIP) was encountered. (*SKIP)
2763     signifies that whatever text was matched leading up to it cannot be part of a
2764     successful match. Consider:
2765 ph10 210 .sp
2766     a+(*SKIP)b
2767     .sp
2768 ph10 211 If the subject is "aaaac...", after the first match attempt fails (starting at
2769 ph10 210 the first character in the string), the starting point skips on to start the
2770 ph10 211 next attempt at "c". Note that a possessive quantifer does not have the same
2771 ph10 456 effect as this example; although it would suppress backtracking during the
2772 ph10 210 first match attempt, the second attempt would start at the second character
2773     instead of skipping on to "c".
2774     .sp
2775 ph10 510 (*SKIP:NAME)
2776 ph10 211 .sp
2777 ph10 512 When (*SKIP) has an associated name, its behaviour is modified. If the
2778     following pattern fails to match, the previous path through the pattern is
2779     searched for the most recent (*MARK) that has the same name. If one is found,
2780     the "bumpalong" advance is to the subject position that corresponds to that
2781     (*MARK) instead of to where (*SKIP) was encountered. If no (*MARK) with a
2782 ph10 716 matching name is found, normal "bumpalong" of one character happens (that is,
2783     the (*SKIP) is ignored).
2784 ph10 510 .sp
2785     (*THEN) or (*THEN:NAME)
2786     .sp
2787 ph10 716 This verb causes a skip to the next innermost alternative if the rest of the
2788     pattern does not match. That is, it cancels pending backtracking, but only
2789     within the current alternative. Its name comes from the observation that it can
2790     be used for a pattern-based if-then-else block:
2791 ph10 210 .sp
2792     ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
2793     .sp
2794 ph10 211 If the COND1 pattern matches, FOO is tried (and possibly further items after
2795 ph10 716 the end of the group if FOO succeeds); on failure, the matcher skips to the
2796 ph10 510 second alternative and tries COND2, without backtracking into COND1. The
2797     behaviour of (*THEN:NAME) is exactly the same as (*MARK:NAME)(*THEN) if the
2798 ph10 716 overall match fails. If (*THEN) is not inside an alternation, it acts like
2799     (*PRUNE).
2800 ph10 551 .P
2801 ph10 716 Note that a subpattern that does not contain a | character is just a part of
2802     the enclosing alternative; it is not a nested alternation with only one
2803     alternative. The effect of (*THEN) extends beyond such a subpattern to the
2804     enclosing alternative. Consider this pattern, where A, B, etc. are complex
2805     pattern fragments that do not contain any | characters at this level:
2806     .sp
2807     A (B(*THEN)C) | D
2808     .sp
2809 ph10 733 If A and B are matched, but there is a failure in C, matching does not
2810 ph10 716 backtrack into A; instead it moves to the next alternative, that is, D.
2811     However, if the subpattern containing (*THEN) is given an alternative, it
2812     behaves differently:
2813     .sp
2814     A (B(*THEN)C | (*FAIL)) | D
2815     .sp
2816     The effect of (*THEN) is now confined to the inner subpattern. After a failure
2817 ph10 733 in C, matching moves to (*FAIL), which causes the whole subpattern to fail
2818     because there are no more alternatives to try. In this case, matching does now
2819 ph10 716 backtrack into A.
2820 ph10 551 .P
2821 ph10 733 Note also that a conditional subpattern is not considered as having two
2822     alternatives, because only one is ever used. In other words, the | character in
2823 ph10 716 a conditional subpattern has a different meaning. Ignoring white space,
2824     consider:
2825 ph10 551 .sp
2826 ph10 716 ^.*? (?(?=a) a | b(*THEN)c )
2827     .sp
2828 ph10 733 If the subject is "ba", this pattern does not match. Because .*? is ungreedy,
2829     it initially matches zero characters. The condition (?=a) then fails, the
2830 ph10 716 character "b" is matched, but "c" is not. At this point, matching does not
2831     backtrack to .*? as might perhaps be expected from the presence of the |
2832     character. The conditional subpattern is part of the single alternative that
2833 ph10 733 comprises the whole pattern, and so the match fails. (If there was a backtrack
2834 ph10 716 into .*?, allowing it to match "b", the match would succeed.)
2835     .P
2836     The verbs just described provide four different "strengths" of control when
2837     subsequent matching fails. (*THEN) is the weakest, carrying on the match at the
2838     next alternative. (*PRUNE) comes next, failing the match at the current
2839     starting position, but allowing an advance to the next character (for an
2840     unanchored pattern). (*SKIP) is similar, except that the advance may be more
2841     than one character. (*COMMIT) is the strongest, causing the entire match to
2842     fail.
2843     .P
2844     If more than one such verb is present in a pattern, the "strongest" one wins.
2845     For example, consider this pattern, where A, B, etc. are complex pattern
2846     fragments:
2847     .sp
2848 ph10 551 (A(*COMMIT)B(*THEN)C|D)
2849     .sp
2850     Once A has matched, PCRE is committed to this match, at the current starting
2851     position. If subsequently B matches, but C does not, the normal (*THEN) action
2852 ph10 716 of trying the next alternative (that is, D) does not happen because (*COMMIT)
2853 ph10 551 overrides.
2854 ph10 210 .
2855 ph10 551 .
2856 nigel 93 .SH "SEE ALSO"
2857     .rs
2858     .sp
2859 ph10 461 \fBpcreapi\fP(3), \fBpcrecallout\fP(3), \fBpcrematching\fP(3),
2860 ph10 456 \fBpcresyntax\fP(3), \fBpcre\fP(3).
2861 ph10 99 .
2862     .
2863     .SH AUTHOR
2864     .rs
2865     .sp
2866     .nf
2867     Philip Hazel
2868     University Computing Service
2869     Cambridge CB2 3QH, England.
2870     .fi
2871     .
2872     .
2873     .SH REVISION
2874     .rs
2875     .sp
2876     .nf
2877 ph10 745 Last updated: 14 November 2011
2878 ph10 607 Copyright (c) 1997-2011 University of Cambridge.
2879 ph10 99 .fi


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

ViewVC Help
Powered by ViewVC 1.1.12