ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 733 - (hide annotations) (download) (as text)
Tue Oct 11 10:29:36 2011 UTC (3 years, 6 months ago) by ph10
File MIME type: text/html
File size: 121852 byte(s)
Source tidies for 8.20-RC3.

1 nigel 63 <html>
2     <head>
3     <title>pcrepattern specification</title>
4     </head>
5     <body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB">
6 nigel 75 <h1>pcrepattern man page</h1>
7     <p>
8     Return to the <a href="index.html">PCRE index page</a>.
9     </p>
10 ph10 111 <p>
11 nigel 75 This page is part of the PCRE HTML documentation. It was generated automatically
12     from the original man page. If there is any nonsense in it, please consult the
13     man page, in case the conversion went wrong.
14 ph10 111 <br>
15 nigel 63 <ul>
16     <li><a name="TOC1" href="#SEC1">PCRE REGULAR EXPRESSION DETAILS</a>
17 ph10 227 <li><a name="TOC2" href="#SEC2">NEWLINE CONVENTIONS</a>
18     <li><a name="TOC3" href="#SEC3">CHARACTERS AND METACHARACTERS</a>
19     <li><a name="TOC4" href="#SEC4">BACKSLASH</a>
20     <li><a name="TOC5" href="#SEC5">CIRCUMFLEX AND DOLLAR</a>
21 ph10 518 <li><a name="TOC6" href="#SEC6">FULL STOP (PERIOD, DOT) AND \N</a>
22 ph10 227 <li><a name="TOC7" href="#SEC7">MATCHING A SINGLE BYTE</a>
23     <li><a name="TOC8" href="#SEC8">SQUARE BRACKETS AND CHARACTER CLASSES</a>
24     <li><a name="TOC9" href="#SEC9">POSIX CHARACTER CLASSES</a>
25     <li><a name="TOC10" href="#SEC10">VERTICAL BAR</a>
26     <li><a name="TOC11" href="#SEC11">INTERNAL OPTION SETTING</a>
27     <li><a name="TOC12" href="#SEC12">SUBPATTERNS</a>
28     <li><a name="TOC13" href="#SEC13">DUPLICATE SUBPATTERN NUMBERS</a>
29     <li><a name="TOC14" href="#SEC14">NAMED SUBPATTERNS</a>
30     <li><a name="TOC15" href="#SEC15">REPETITION</a>
32     <li><a name="TOC17" href="#SEC17">BACK REFERENCES</a>
33     <li><a name="TOC18" href="#SEC18">ASSERTIONS</a>
34     <li><a name="TOC19" href="#SEC19">CONDITIONAL SUBPATTERNS</a>
35     <li><a name="TOC20" href="#SEC20">COMMENTS</a>
36     <li><a name="TOC21" href="#SEC21">RECURSIVE PATTERNS</a>
37     <li><a name="TOC22" href="#SEC22">SUBPATTERNS AS SUBROUTINES</a>
38 ph10 345 <li><a name="TOC23" href="#SEC23">ONIGURUMA SUBROUTINE SYNTAX</a>
39     <li><a name="TOC24" href="#SEC24">CALLOUTS</a>
40     <li><a name="TOC25" href="#SEC25">BACKTRACKING CONTROL</a>
41     <li><a name="TOC26" href="#SEC26">SEE ALSO</a>
42     <li><a name="TOC27" href="#SEC27">AUTHOR</a>
43     <li><a name="TOC28" href="#SEC28">REVISION</a>
44 nigel 63 </ul>
45     <br><a name="SEC1" href="#TOC1">PCRE REGULAR EXPRESSION DETAILS</a><br>
46     <P>
47 ph10 208 The syntax and semantics of the regular expressions that are supported by PCRE
48     are described in detail below. There is a quick-reference syntax summary in the
49     <a href="pcresyntax.html"><b>pcresyntax</b></a>
50 ph10 345 page. PCRE tries to match Perl syntax and semantics as closely as it can. PCRE
51     also supports some alternative regular expression syntax (which does not
52     conflict with the Perl syntax) in order to provide some compatibility with
53     regular expressions in Python, .NET, and Oniguruma.
54     </P>
55     <P>
56     Perl's regular expressions are described in its own documentation, and
57 ph10 208 regular expressions in general are covered in a number of books, some of which
58     have copious examples. Jeffrey Friedl's "Mastering Regular Expressions",
59     published by O'Reilly, covers regular expressions in great detail. This
60     description of PCRE's regular expressions is intended as reference material.
61 nigel 63 </P>
62     <P>
63 nigel 75 The original operation of PCRE was on strings of one-byte characters. However,
64 ph10 461 there is now also support for UTF-8 character strings. To use this,
65     PCRE must be built to include UTF-8 support, and you must call
66     <b>pcre_compile()</b> or <b>pcre_compile2()</b> with the PCRE_UTF8 option. There
67     is also a special sequence that can be given at the start of a pattern:
68 ph10 416 <pre>
69     (*UTF8)
70     </pre>
71     Starting a pattern with this sequence is equivalent to setting the PCRE_UTF8
72     option. This feature is not Perl-compatible. How setting UTF-8 mode affects
73     pattern matching is mentioned in several places below. There is also a summary
74     of UTF-8 features in the
75 ph10 678 <a href="pcreunicode.html"><b>pcreunicode</b></a>
76 nigel 63 page.
77     </P>
78     <P>
79 ph10 535 Another special sequence that may appear at the start of a pattern or in
80     combination with (*UTF8) is:
81     <pre>
82     (*UCP)
83     </pre>
84     This has the same effect as setting the PCRE_UCP option: it causes sequences
85     such as \d and \w to use Unicode properties to determine character types,
86     instead of recognizing only characters with codes less than 128 via a lookup
87     table.
88     </P>
89     <P>
90 ph10 579 If a pattern starts with (*NO_START_OPT), it has the same effect as setting the
91     PCRE_NO_START_OPTIMIZE option either at compile or matching time. There are
92     also some more of these special sequences that are concerned with the handling
93     of newlines; they are described below.
94     </P>
95     <P>
96 nigel 77 The remainder of this document discusses the patterns that are supported by
97     PCRE when its main matching function, <b>pcre_exec()</b>, is used.
98     From release 6.0, PCRE offers a second matching function,
99     <b>pcre_dfa_exec()</b>, which matches using a different algorithm that is not
100 ph10 172 Perl-compatible. Some of the features discussed below are not available when
101     <b>pcre_dfa_exec()</b> is used. The advantages and disadvantages of the
102     alternative function, and how it differs from the normal function, are
103     discussed in the
104 nigel 77 <a href="pcrematching.html"><b>pcrematching</b></a>
105     page.
106 ph10 567 <a name="newlines"></a></P>
107 ph10 227 <br><a name="SEC2" href="#TOC1">NEWLINE CONVENTIONS</a><br>
108 nigel 77 <P>
109 ph10 227 PCRE supports five different conventions for indicating line breaks in
110     strings: a single CR (carriage return) character, a single LF (linefeed)
111     character, the two-character sequence CRLF, any of the three preceding, or any
112     Unicode newline sequence. The
113     <a href="pcreapi.html"><b>pcreapi</b></a>
114     page has
115     <a href="pcreapi.html#newlines">further discussion</a>
116     about newlines, and shows how to set the newline convention in the
117     <i>options</i> arguments for the compiling and matching functions.
118     </P>
119     <P>
120     It is also possible to specify a newline convention by starting a pattern
121     string with one of the following five sequences:
122     <pre>
123     (*CR) carriage return
124     (*LF) linefeed
125     (*CRLF) carriage return, followed by linefeed
126     (*ANYCRLF) any of the three above
127     (*ANY) all Unicode newline sequences
128     </pre>
129 ph10 461 These override the default and the options given to <b>pcre_compile()</b> or
130     <b>pcre_compile2()</b>. For example, on a Unix system where LF is the default
131     newline sequence, the pattern
132 ph10 227 <pre>
133     (*CR)a.b
134     </pre>
135     changes the convention to CR. That pattern matches "a\nb" because LF is no
136     longer a newline. Note that these special settings, which are not
137     Perl-compatible, are recognized only at the very start of a pattern, and that
138 ph10 231 they must be in upper case. If more than one of them is present, the last one
139     is used.
140 ph10 227 </P>
141 ph10 231 <P>
142 ph10 518 The newline convention affects the interpretation of the dot metacharacter when
143     PCRE_DOTALL is not set, and also the behaviour of \N. However, it does not
144     affect what the \R escape sequence matches. By default, this is any Unicode
145     newline sequence, for Perl compatibility. However, this can be changed; see the
146     description of \R in the section entitled
147 ph10 231 <a href="#newlineseq">"Newline sequences"</a>
148 ph10 247 below. A change of \R setting can be combined with a change of newline
149     convention.
150 ph10 231 </P>
151 ph10 227 <br><a name="SEC3" href="#TOC1">CHARACTERS AND METACHARACTERS</a><br>
152     <P>
153 nigel 63 A regular expression is a pattern that is matched against a subject string from
154     left to right. Most characters stand for themselves in a pattern, and match the
155     corresponding characters in the subject. As a trivial example, the pattern
156     <pre>
157     The quick brown fox
158 nigel 75 </pre>
159 nigel 77 matches a portion of a subject string that is identical to itself. When
160     caseless matching is specified (the PCRE_CASELESS option), letters are matched
161     independently of case. In UTF-8 mode, PCRE always understands the concept of
162     case for characters whose values are less than 128, so caseless matching is
163     always possible. For characters with higher values, the concept of case is
164     supported if PCRE is compiled with Unicode property support, but not otherwise.
165     If you want to use caseless matching for characters 128 and above, you must
166     ensure that PCRE is compiled with Unicode property support as well as with
167     UTF-8 support.
168     </P>
169     <P>
170     The power of regular expressions comes from the ability to include alternatives
171     and repetitions in the pattern. These are encoded in the pattern by the use of
172 nigel 75 <i>metacharacters</i>, which do not stand for themselves but instead are
173 nigel 63 interpreted in some special way.
174     </P>
175     <P>
176 nigel 75 There are two different sets of metacharacters: those that are recognized
177 nigel 63 anywhere in the pattern except within square brackets, and those that are
178 nigel 93 recognized within square brackets. Outside square brackets, the metacharacters
179     are as follows:
180 nigel 63 <pre>
181     \ general escape character with several uses
182     ^ assert start of string (or line, in multiline mode)
183     $ assert end of string (or line, in multiline mode)
184     . match any character except newline (by default)
185     [ start character class definition
186     | start of alternative branch
187     ( start subpattern
188     ) end subpattern
189     ? extends the meaning of (
190     also 0 or 1 quantifier
191     also quantifier minimizer
192     * 0 or more quantifier
193     + 1 or more quantifier
194     also "possessive quantifier"
195     { start min/max quantifier
196 nigel 75 </pre>
197 nigel 63 Part of a pattern that is in square brackets is called a "character class". In
198 nigel 75 a character class the only metacharacters are:
199 nigel 63 <pre>
200     \ general escape character
201     ^ negate the class, but only if the first character
202     - indicates character range
203 nigel 75 [ POSIX character class (only if followed by POSIX syntax)
204 nigel 63 ] terminates the character class
205 nigel 75 </pre>
206     The following sections describe the use of each of the metacharacters.
207 nigel 63 </P>
208 ph10 227 <br><a name="SEC4" href="#TOC1">BACKSLASH</a><br>
209 nigel 63 <P>
210     The backslash character has several uses. Firstly, if it is followed by a
211 ph10 579 character that is not a number or a letter, it takes away any special meaning
212     that character may have. This use of backslash as an escape character applies
213     both inside and outside character classes.
214 nigel 63 </P>
215     <P>
216     For example, if you want to match a * character, you write \* in the pattern.
217     This escaping action applies whether or not the following character would
218 nigel 75 otherwise be interpreted as a metacharacter, so it is always safe to precede a
219     non-alphanumeric with backslash to specify that it stands for itself. In
220 nigel 63 particular, if you want to match a backslash, you write \\.
221     </P>
222     <P>
223 ph10 579 In UTF-8 mode, only ASCII numbers and letters have any special meaning after a
224     backslash. All other characters (in particular, those whose codepoints are
225     greater than 127) are treated as literals.
226     </P>
227     <P>
228 nigel 63 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
229     pattern (other than in a character class) and characters between a # outside
230 nigel 91 a character class and the next newline are ignored. An escaping backslash can
231     be used to include a whitespace or # character as part of the pattern.
232 nigel 63 </P>
233     <P>
234     If you want to remove the special meaning from a sequence of characters, you
235     can do so by putting them between \Q and \E. This is different from Perl in
236     that $ and @ are handled as literals in \Q...\E sequences in PCRE, whereas in
237     Perl, $ and @ cause variable interpolation. Note the following examples:
238     <pre>
239     Pattern PCRE matches Perl matches
240 nigel 75
241     \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
242 nigel 63 \Qabc\$xyz\E abc\$xyz abc\$xyz
243     \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
244 nigel 75 </pre>
245 nigel 63 The \Q...\E sequence is recognized both inside and outside character classes.
246 ph10 654 An isolated \E that is not preceded by \Q is ignored. If \Q is not followed
247     by \E later in the pattern, the literal interpretation continues to the end of
248     the pattern (that is, \E is assumed at the end). If the isolated \Q is inside
249     a character class, this causes an error, because the character class is not
250     terminated.
251 nigel 75 <a name="digitsafterbackslash"></a></P>
252     <br><b>
253     Non-printing characters
254     </b><br>
255 nigel 63 <P>
256     A second use of backslash provides a way of encoding non-printing characters
257     in patterns in a visible manner. There is no restriction on the appearance of
258     non-printing characters, apart from the binary zero that terminates a pattern,
259 ph10 461 but when a pattern is being prepared by text editing, it is often easier to use
260     one of the following escape sequences than the binary character it represents:
261 nigel 63 <pre>
262     \a alarm, that is, the BEL character (hex 07)
263 ph10 579 \cx "control-x", where x is any ASCII character
264 nigel 63 \e escape (hex 1B)
265     \f formfeed (hex 0C)
266 ph10 227 \n linefeed (hex 0A)
267 nigel 63 \r carriage return (hex 0D)
268     \t tab (hex 09)
269 ph10 488 \ddd character with octal code ddd, or back reference
270 nigel 63 \xhh character with hex code hh
271 nigel 87 \x{hhh..} character with hex code hhh..
272 nigel 75 </pre>
273 nigel 63 The precise effect of \cx is as follows: if x is a lower case letter, it
274     is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
275 ph10 579 Thus \cz becomes hex 1A (z is 7A), but \c{ becomes hex 3B ({ is 7B), while
276     \c; becomes hex 7B (; is 3B). If the byte following \c has a value greater
277     than 127, a compile-time error occurs. This locks out non-ASCII characters in
278     both byte mode and UTF-8 mode. (When PCRE is compiled in EBCDIC mode, all byte
279     values are valid. A lower case letter is converted to upper case, and then the
280     0xc0 bits are flipped.)
281 nigel 63 </P>
282     <P>
283     After \x, from zero to two hexadecimal digits are read (letters can be in
284 nigel 87 upper or lower case). Any number of hexadecimal digits may appear between \x{
285     and }, but the value of the character code must be less than 256 in non-UTF-8
286 ph10 211 mode, and less than 2**31 in UTF-8 mode. That is, the maximum value in
287     hexadecimal is 7FFFFFFF. Note that this is bigger than the largest Unicode code
288     point, which is 10FFFF.
289 nigel 63 </P>
290     <P>
291 ph10 211 If characters other than hexadecimal digits appear between \x{ and }, or if
292     there is no terminating }, this form of escape is not recognized. Instead, the
293     initial \x will be interpreted as a basic hexadecimal escape, with no
294     following digits, giving a character whose value is zero.
295     </P>
296     <P>
297 nigel 63 Characters whose value is less than 256 can be defined by either of the two
298 nigel 87 syntaxes for \x. There is no difference in the way they are handled. For
299     example, \xdc is exactly the same as \x{dc}.
300 nigel 63 </P>
301     <P>
302 nigel 91 After \0 up to two further octal digits are read. If there are fewer than two
303     digits, just those that are present are used. Thus the sequence \0\x\07
304     specifies two binary zeros followed by a BEL character (code value 7). Make
305     sure you supply two digits after the initial zero if the pattern character that
306     follows is itself an octal digit.
307 nigel 63 </P>
308     <P>
309     The handling of a backslash followed by a digit other than 0 is complicated.
310     Outside a character class, PCRE reads it and any following digits as a decimal
311     number. If the number is less than 10, or if there have been at least that many
312     previous capturing left parentheses in the expression, the entire sequence is
313     taken as a <i>back reference</i>. A description of how this works is given
314 nigel 75 <a href="#backreferences">later,</a>
315     following the discussion of
316     <a href="#subpattern">parenthesized subpatterns.</a>
317 nigel 63 </P>
318     <P>
319     Inside a character class, or if the decimal number is greater than 9 and there
320     have not been that many capturing subpatterns, PCRE re-reads up to three octal
321 nigel 93 digits following the backslash, and uses them to generate a data character. Any
322 nigel 91 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
323     character specified in octal must be less than \400. In UTF-8 mode, values up
324     to \777 are permitted. For example:
325 nigel 63 <pre>
326     \040 is another way of writing a space
327 nigel 75 \40 is the same, provided there are fewer than 40 previous capturing subpatterns
328 nigel 63 \7 is always a back reference
329 nigel 75 \11 might be a back reference, or another way of writing a tab
330 nigel 63 \011 is always a tab
331     \0113 is a tab followed by the character "3"
332 nigel 75 \113 might be a back reference, otherwise the character with octal code 113
333     \377 might be a back reference, otherwise the byte consisting entirely of 1 bits
334     \81 is either a back reference, or a binary zero followed by the two characters "8" and "1"
335     </pre>
336 nigel 63 Note that octal values of 100 or greater must not be introduced by a leading
337     zero, because no more than three octal digits are ever read.
338     </P>
339     <P>
340 nigel 91 All the sequences that define a single character value can be used both inside
341     and outside character classes. In addition, inside a character class, the
342 ph10 518 sequence \b is interpreted as the backspace character (hex 08). The sequences
343     \B, \N, \R, and \X are not special inside a character class. Like any other
344     unrecognized escape sequences, they are treated as the literal characters "B",
345     "N", "R", and "X" by default, but cause an error if the PCRE_EXTRA option is
346     set. Outside a character class, these sequences have different meanings.
347 nigel 63 </P>
348 nigel 75 <br><b>
349 nigel 93 Absolute and relative back references
350     </b><br>
351     <P>
352 ph10 208 The sequence \g 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 \g{name}. Back references are discussed
355 nigel 93 <a href="#backreferences">later,</a>
356     following the discussion of
357     <a href="#subpattern">parenthesized subpatterns.</a>
358     </P>
359     <br><b>
360 ph10 345 Absolute and relative subroutine calls
361     </b><br>
362     <P>
363     For compatibility with Oniguruma, the non-Perl syntax \g followed by a name or
364     a number enclosed either in angle brackets or single quotes, is an alternative
365     syntax for referencing a subpattern as a "subroutine". Details are discussed
366     <a href="#onigurumasubroutines">later.</a>
367     Note that \g{...} (Perl syntax) and \g&#60;...&#62; (Oniguruma syntax) are <i>not</i>
368 ph10 461 synonymous. The former is a back reference; the latter is a
369 ph10 454 <a href="#subpatternsassubroutines">subroutine</a>
370     call.
371 ph10 518 <a name="genericchartypes"></a></P>
372 ph10 345 <br><b>
373 nigel 75 Generic character types
374     </b><br>
375 nigel 63 <P>
376 ph10 518 Another use of backslash is for specifying generic character types:
377 nigel 63 <pre>
378     \d any decimal digit
379     \D any character that is not a decimal digit
380 ph10 182 \h any horizontal whitespace character
381     \H any character that is not a horizontal whitespace character
382 nigel 63 \s any whitespace character
383     \S any character that is not a whitespace character
384 ph10 182 \v any vertical whitespace character
385     \V any character that is not a vertical whitespace character
386 nigel 63 \w any "word" character
387     \W any "non-word" character
388 nigel 75 </pre>
389 ph10 535 There is also the single sequence \N, which matches a non-newline character.
390     This is the same as
391 ph10 518 <a href="#fullstopdot">the "." metacharacter</a>
392     when PCRE_DOTALL is not set.
393 nigel 63 </P>
394     <P>
395 ph10 518 Each pair of lower and upper case escape sequences partitions the complete set
396     of characters into two disjoint sets. Any given character matches one, and only
397 ph10 535 one, of each pair. The sequences can appear both inside and outside character
398 nigel 75 classes. They each match one character of the appropriate type. If the current
399 ph10 535 matching point is at the end of the subject string, all of them fail, because
400 nigel 75 there is no character to match.
401 nigel 63 </P>
402     <P>
403     For compatibility with Perl, \s does not match the VT character (code 11).
404     This makes it different from the the POSIX "space" class. The \s characters
405 ph10 182 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
406 nigel 91 included in a Perl script, \s may match the VT character. In PCRE, it never
407 ph10 182 does.
408 nigel 63 </P>
409     <P>
410 ph10 535 A "word" character is an underscore or any character that is a letter or digit.
411     By default, the definition of letters and digits is controlled by PCRE's
412     low-valued character tables, and may vary if locale-specific matching is taking
413     place (see
414     <a href="pcreapi.html#localesupport">"Locale support"</a>
415     in the
416     <a href="pcreapi.html"><b>pcreapi</b></a>
417     page). For example, in a French locale such as "fr_FR" in Unix-like systems,
418     or "french" in Windows, some character codes greater than 128 are used for
419     accented letters, and these are then matched by \w. The use of locales with
420     Unicode is discouraged.
421 ph10 182 </P>
422     <P>
423 ph10 535 By default, in UTF-8 mode, characters with values greater than 128 never match
424     \d, \s, or \w, and always match \D, \S, and \W. These sequences retain
425     their original meanings from before UTF-8 support was available, mainly for
426     efficiency reasons. However, if PCRE is compiled with Unicode property support,
427     and the PCRE_UCP option is set, the behaviour is changed so that Unicode
428     properties are used to determine character types, as follows:
429     <pre>
430     \d any character that \p{Nd} matches (decimal digit)
431     \s any character that \p{Z} matches, plus HT, LF, FF, CR
432     \w any character that \p{L} or \p{N} matches, plus underscore
433     </pre>
434     The upper case escapes match the inverse sets of characters. Note that \d
435     matches only decimal digits, whereas \w matches any Unicode digit, as well as
436     any Unicode letter, and underscore. Note also that PCRE_UCP affects \b, and
437     \B because they are defined in terms of \w and \W. Matching these sequences
438     is noticeably slower when PCRE_UCP is set.
439     </P>
440     <P>
441 ph10 579 The sequences \h, \H, \v, and \V are features that were added to Perl at
442 ph10 572 release 5.10. In contrast to the other sequences, which match only ASCII
443     characters by default, these always match certain high-valued codepoints in
444     UTF-8 mode, whether or not PCRE_UCP is set. The horizontal space characters
445     are:
446 ph10 182 <pre>
447     U+0009 Horizontal tab
448     U+0020 Space
449     U+00A0 Non-break space
450     U+1680 Ogham space mark
451     U+180E Mongolian vowel separator
452     U+2000 En quad
453     U+2001 Em quad
454     U+2002 En space
455     U+2003 Em space
456     U+2004 Three-per-em space
457     U+2005 Four-per-em space
458     U+2006 Six-per-em space
459     U+2007 Figure space
460     U+2008 Punctuation space
461     U+2009 Thin space
462     U+200A Hair space
463     U+202F Narrow no-break space
464     U+205F Medium mathematical space
465     U+3000 Ideographic space
466     </pre>
467     The vertical space characters are:
468     <pre>
469     U+000A Linefeed
470     U+000B Vertical tab
471     U+000C Formfeed
472     U+000D Carriage return
473     U+0085 Next line
474     U+2028 Line separator
475     U+2029 Paragraph separator
476 ph10 535 <a name="newlineseq"></a></PRE>
477 ph10 182 </P>
478 nigel 93 <br><b>
479     Newline sequences
480     </b><br>
481     <P>
482 ph10 231 Outside a character class, by default, the escape sequence \R matches any
483 ph10 572 Unicode newline sequence. In non-UTF-8 mode \R is equivalent to the following:
484 nigel 93 <pre>
485     (?&#62;\r\n|\n|\x0b|\f|\r|\x85)
486     </pre>
487     This is an example of an "atomic group", details of which are given
488     <a href="#atomicgroup">below.</a>
489     This particular group matches either the two-character sequence CR followed by
490     LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
491     U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
492     line, U+0085). The two-character sequence is treated as a single unit that
493     cannot be split.
494     </P>
495     <P>
496     In UTF-8 mode, two additional characters whose codepoints are greater than 255
497     are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
498     Unicode character property support is not needed for these characters to be
499     recognized.
500     </P>
501     <P>
502 ph10 231 It is possible to restrict \R to match only CR, LF, or CRLF (instead of the
503     complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
504 ph10 247 either at compile time or when the pattern is matched. (BSR is an abbrevation
505     for "backslash R".) This can be made the default when PCRE is built; if this is
506     the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
507     It is also possible to specify these settings by starting a pattern string with
508     one of the following sequences:
509 ph10 231 <pre>
510     (*BSR_ANYCRLF) CR, LF, or CRLF only
511     (*BSR_UNICODE) any Unicode newline sequence
512     </pre>
513 ph10 461 These override the default and the options given to <b>pcre_compile()</b> or
514     <b>pcre_compile2()</b>, but they can be overridden by options given to
515     <b>pcre_exec()</b> or <b>pcre_dfa_exec()</b>. Note that these special settings,
516     which are not Perl-compatible, are recognized only at the very start of a
517     pattern, and that they must be in upper case. If more than one of them is
518     present, the last one is used. They can be combined with a change of newline
519 ph10 535 convention; for example, a pattern can start with:
520 ph10 247 <pre>
521     (*ANY)(*BSR_ANYCRLF)
522     </pre>
523 ph10 535 They can also be combined with the (*UTF8) or (*UCP) special sequences. Inside
524     a character class, \R is treated as an unrecognized escape sequence, and so
525     matches the letter "R" by default, but causes an error if PCRE_EXTRA is set.
526 nigel 75 <a name="uniextseq"></a></P>
527     <br><b>
528     Unicode character properties
529     </b><br>
530     <P>
531     When PCRE is built with Unicode character property support, three additional
532 ph10 185 escape sequences that match characters with specific properties are available.
533     When not in UTF-8 mode, these sequences are of course limited to testing
534     characters whose codepoints are less than 256, but they do work in this mode.
535     The extra escape sequences are:
536 nigel 75 <pre>
537 nigel 87 \p{<i>xx</i>} a character with the <i>xx</i> property
538     \P{<i>xx</i>} a character without the <i>xx</i> property
539     \X an extended Unicode sequence
540 nigel 75 </pre>
541 nigel 87 The property names represented by <i>xx</i> above are limited to the Unicode
542 ph10 518 script names, the general category properties, "Any", which matches any
543     character (including newline), and some special PCRE properties (described
544 ph10 535 in the
545 ph10 518 <a href="#extraprops">next section).</a>
546     Other Perl properties such as "InMusicalSymbols" are not currently supported by
547     PCRE. Note that \P{Any} does not match any characters, so always causes a
548     match failure.
549 nigel 63 </P>
550     <P>
551 nigel 87 Sets of Unicode characters are defined as belonging to certain scripts. A
552     character from one of these sets can be matched using a script name. For
553     example:
554 nigel 75 <pre>
555 nigel 87 \p{Greek}
556     \P{Han}
557     </pre>
558     Those that are not part of an identified script are lumped together as
559     "Common". The current list of scripts is:
560     </P>
561     <P>
562     Arabic,
563     Armenian,
564 ph10 507 Avestan,
565 nigel 93 Balinese,
566 ph10 507 Bamum,
567 nigel 87 Bengali,
568     Bopomofo,
569     Braille,
570     Buginese,
571     Buhid,
572     Canadian_Aboriginal,
573 ph10 507 Carian,
574     Cham,
575 nigel 87 Cherokee,
576     Common,
577     Coptic,
578 nigel 93 Cuneiform,
579 nigel 87 Cypriot,
580     Cyrillic,
581     Deseret,
582     Devanagari,
583 ph10 507 Egyptian_Hieroglyphs,
584 nigel 87 Ethiopic,
585     Georgian,
586     Glagolitic,
587     Gothic,
588     Greek,
589     Gujarati,
590     Gurmukhi,
591     Han,
592     Hangul,
593     Hanunoo,
594     Hebrew,
595     Hiragana,
596 ph10 507 Imperial_Aramaic,
597 nigel 87 Inherited,
598 ph10 507 Inscriptional_Pahlavi,
599     Inscriptional_Parthian,
600     Javanese,
601     Kaithi,
602 nigel 87 Kannada,
603     Katakana,
604 ph10 507 Kayah_Li,
605 nigel 87 Kharoshthi,
606     Khmer,
607     Lao,
608     Latin,
609 ph10 507 Lepcha,
610 nigel 87 Limbu,
611     Linear_B,
612 ph10 507 Lisu,
613     Lycian,
614     Lydian,
615 nigel 87 Malayalam,
616 ph10 507 Meetei_Mayek,
617 nigel 87 Mongolian,
618     Myanmar,
619     New_Tai_Lue,
620 nigel 93 Nko,
621 nigel 87 Ogham,
622     Old_Italic,
623     Old_Persian,
624 ph10 507 Old_South_Arabian,
625     Old_Turkic,
626     Ol_Chiki,
627 nigel 87 Oriya,
628     Osmanya,
629 nigel 93 Phags_Pa,
630     Phoenician,
631 ph10 507 Rejang,
632 nigel 87 Runic,
633 ph10 507 Samaritan,
634     Saurashtra,
635 nigel 87 Shavian,
636     Sinhala,
637 ph10 507 Sundanese,
638 nigel 87 Syloti_Nagri,
639     Syriac,
640     Tagalog,
641     Tagbanwa,
642     Tai_Le,
643 ph10 507 Tai_Tham,
644     Tai_Viet,
645 nigel 87 Tamil,
646     Telugu,
647     Thaana,
648     Thai,
649     Tibetan,
650     Tifinagh,
651     Ugaritic,
652 ph10 507 Vai,
653 nigel 87 Yi.
654     </P>
655     <P>
656 ph10 518 Each character has exactly one Unicode general category property, specified by
657     a two-letter abbreviation. For compatibility with Perl, negation can be
658     specified by including a circumflex between the opening brace and the property
659     name. For example, \p{^Lu} is the same as \P{Lu}.
660 nigel 87 </P>
661     <P>
662     If only one letter is specified with \p or \P, it includes all the general
663     category properties that start with that letter. In this case, in the absence
664     of negation, the curly brackets in the escape sequence are optional; these two
665     examples have the same effect:
666     <pre>
667 nigel 75 \p{L}
668     \pL
669     </pre>
670 nigel 87 The following general category property codes are supported:
671 nigel 75 <pre>
672     C Other
673     Cc Control
674     Cf Format
675     Cn Unassigned
676     Co Private use
677     Cs Surrogate
679     L Letter
680     Ll Lower case letter
681     Lm Modifier letter
682     Lo Other letter
683     Lt Title case letter
684     Lu Upper case letter
686     M Mark
687     Mc Spacing mark
688     Me Enclosing mark
689     Mn Non-spacing mark
691     N Number
692     Nd Decimal number
693     Nl Letter number
694     No Other number
696     P Punctuation
697     Pc Connector punctuation
698     Pd Dash punctuation
699     Pe Close punctuation
700     Pf Final punctuation
701     Pi Initial punctuation
702     Po Other punctuation
703     Ps Open punctuation
705     S Symbol
706     Sc Currency symbol
707     Sk Modifier symbol
708     Sm Mathematical symbol
709     So Other symbol
711     Z Separator
712     Zl Line separator
713     Zp Paragraph separator
714     Zs Space separator
715     </pre>
716 nigel 87 The special property L& is also supported: it matches a character that has
717     the Lu, Ll, or Lt property, in other words, a letter that is not classified as
718     a modifier or "other".
719 nigel 75 </P>
720     <P>
721 ph10 211 The Cs (Surrogate) property applies only to characters in the range U+D800 to
722     U+DFFF. Such characters are not valid in UTF-8 strings (see RFC 3629) and so
723     cannot be tested by PCRE, unless UTF-8 validity checking has been turned off
724     (see the discussion of PCRE_NO_UTF8_CHECK in the
725     <a href="pcreapi.html"><b>pcreapi</b></a>
726 ph10 453 page). Perl does not support the Cs property.
727 ph10 211 </P>
728     <P>
729 ph10 453 The long synonyms for property names that Perl supports (such as \p{Letter})
730 nigel 91 are not supported by PCRE, nor is it permitted to prefix any of these
731 nigel 87 properties with "Is".
732     </P>
733     <P>
734     No character that is in the Unicode table has the Cn (unassigned) property.
735     Instead, this property is assumed for any code point that is not in the
736     Unicode table.
737     </P>
738     <P>
739 nigel 75 Specifying caseless matching does not affect these escape sequences. For
740     example, \p{Lu} always matches only upper case letters.
741     </P>
742     <P>
743     The \X escape matches any number of Unicode characters that form an extended
744     Unicode sequence. \X is equivalent to
745     <pre>
746     (?&#62;\PM\pM*)
747     </pre>
748     That is, it matches a character without the "mark" property, followed by zero
749     or more characters with the "mark" property, and treats the sequence as an
750     atomic group
751     <a href="#atomicgroup">(see below).</a>
752     Characters with the "mark" property are typically accents that affect the
753 ph10 185 preceding character. None of them have codepoints less than 256, so in
754     non-UTF-8 mode \X matches any one character.
755 nigel 75 </P>
756     <P>
757 ph10 654 Note that recent versions of Perl have changed \X to match what Unicode calls
758     an "extended grapheme cluster", which has a more complicated definition.
759     </P>
760     <P>
761 nigel 75 Matching characters by Unicode property is not fast, because PCRE has to search
762     a structure that contains data for over fifteen thousand characters. That is
763     why the traditional escape sequences such as \d and \w do not use Unicode
764 ph10 535 properties in PCRE by default, though you can make them do so by setting the
765     PCRE_UCP option for <b>pcre_compile()</b> or by starting the pattern with
766     (*UCP).
767 ph10 518 <a name="extraprops"></a></P>
768     <br><b>
769     PCRE's additional properties
770     </b><br>
771     <P>
772 ph10 535 As well as the standard Unicode properties described in the previous
773     section, PCRE supports four more that make it possible to convert traditional
774 ph10 518 escape sequences such as \w and \s and POSIX character classes to use Unicode
775 ph10 535 properties. PCRE uses these non-standard, non-Perl properties internally when
776     PCRE_UCP is set. They are:
777 ph10 518 <pre>
778     Xan Any alphanumeric character
779     Xps Any POSIX space character
780     Xsp Any Perl space character
781     Xwd Any Perl "word" character
782     </pre>
783 ph10 535 Xan matches characters that have either the L (letter) or the N (number)
784     property. Xps matches the characters tab, linefeed, vertical tab, formfeed, or
785 ph10 518 carriage return, and any other character that has the Z (separator) property.
786 ph10 535 Xsp is the same as Xps, except that vertical tab is excluded. Xwd matches the
787 ph10 518 same characters as Xan, plus underscore.
788 ph10 172 <a name="resetmatchstart"></a></P>
789     <br><b>
790     Resetting the match start
791     </b><br>
792     <P>
793 ph10 572 The escape sequence \K causes any previously matched characters not to be
794     included in the final matched sequence. For example, the pattern:
795 ph10 172 <pre>
796     foo\Kbar
797     </pre>
798     matches "foobar", but reports that it has matched "bar". This feature is
799     similar to a lookbehind assertion
800     <a href="#lookbehind">(described below).</a>
801     However, in this case, the part of the subject before the real match does not
802     have to be of fixed length, as lookbehind assertions do. The use of \K does
803     not interfere with the setting of
804     <a href="#subpattern">captured substrings.</a>
805     For example, when the pattern
806     <pre>
807     (foo)\Kbar
808     </pre>
809     matches "foobar", the first substring is still set to "foo".
810 ph10 507 </P>
811     <P>
812     Perl documents that the use of \K within assertions is "not well defined". In
813     PCRE, \K is acted upon when it occurs inside positive assertions, but is
814     ignored in negative assertions.
815 nigel 75 <a name="smallassertions"></a></P>
816     <br><b>
817     Simple assertions
818     </b><br>
819     <P>
820 nigel 93 The final use of backslash is for certain simple assertions. An assertion
821 nigel 63 specifies a condition that has to be met at a particular point in a match,
822     without consuming any characters from the subject string. The use of
823 nigel 75 subpatterns for more complicated assertions is described
824     <a href="#bigassertions">below.</a>
825 nigel 91 The backslashed assertions are:
826 nigel 63 <pre>
827     \b matches at a word boundary
828     \B matches when not at a word boundary
829 nigel 93 \A matches at the start of the subject
830     \Z matches at the end of the subject
831     also matches before a newline at the end of the subject
832     \z matches only at the end of the subject
833     \G matches at the first matching position in the subject
834 nigel 75 </pre>
835 ph10 518 Inside a character class, \b has a different meaning; it matches the backspace
836 ph10 535 character. If any other of these assertions appears in a character class, by
837 ph10 518 default it matches the corresponding literal character (for example, \B
838     matches the letter B). However, if the PCRE_EXTRA option is set, an "invalid
839     escape sequence" error is generated instead.
840 nigel 63 </P>
841     <P>
842     A word boundary is a position in the subject string where the current character
843     and the previous character do not both match \w or \W (i.e. one matches
844     \w and the other matches \W), or the start or end of the string if the
845 ph10 535 first or last character matches \w, respectively. In UTF-8 mode, the meanings
846     of \w and \W can be changed by setting the PCRE_UCP option. When this is
847     done, it also affects \b and \B. Neither PCRE nor Perl has a separate "start
848     of word" or "end of word" metasequence. However, whatever follows \b normally
849     determines which it is. For example, the fragment \ba matches "a" at the start
850     of a word.
851 nigel 63 </P>
852     <P>
853     The \A, \Z, and \z assertions differ from the traditional circumflex and
854 nigel 75 dollar (described in the next section) in that they only ever match at the very
855     start and end of the subject string, whatever options are set. Thus, they are
856     independent of multiline mode. These three assertions are not affected by the
857     PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
858     circumflex and dollar metacharacters. However, if the <i>startoffset</i>
859     argument of <b>pcre_exec()</b> is non-zero, indicating that matching is to start
860     at a point other than the beginning of the subject, \A can never match. The
861 nigel 91 difference between \Z and \z is that \Z matches before a newline at the end
862     of the string as well as at the very end, whereas \z matches only at the end.
863 nigel 63 </P>
864     <P>
865     The \G assertion is true only when the current matching position is at the
866     start point of the match, as specified by the <i>startoffset</i> argument of
867     <b>pcre_exec()</b>. It differs from \A when the value of <i>startoffset</i> is
868     non-zero. By calling <b>pcre_exec()</b> multiple times with appropriate
869     arguments, you can mimic Perl's /g option, and it is in this kind of
870     implementation where \G can be useful.
871     </P>
872     <P>
873     Note, however, that PCRE's interpretation of \G, as the start of the current
874     match, is subtly different from Perl's, which defines it as the end of the
875     previous match. In Perl, these can be different when the previously matched
876     string was empty. Because PCRE does just one match at a time, it cannot
877     reproduce this behaviour.
878     </P>
879     <P>
880     If all the alternatives of a pattern begin with \G, the expression is anchored
881     to the starting match position, and the "anchored" flag is set in the compiled
882     regular expression.
883     </P>
884 ph10 227 <br><a name="SEC5" href="#TOC1">CIRCUMFLEX AND DOLLAR</a><br>
885 nigel 63 <P>
886     Outside a character class, in the default matching mode, the circumflex
887 nigel 75 character is an assertion that is true only if the current matching point is
888 nigel 63 at the start of the subject string. If the <i>startoffset</i> argument of
889     <b>pcre_exec()</b> is non-zero, circumflex can never match if the PCRE_MULTILINE
890     option is unset. Inside a character class, circumflex has an entirely different
891 nigel 75 meaning
892     <a href="#characterclass">(see below).</a>
893 nigel 63 </P>
894     <P>
895     Circumflex need not be the first character of the pattern if a number of
896     alternatives are involved, but it should be the first thing in each alternative
897     in which it appears if the pattern is ever to match that branch. If all
898     possible alternatives start with a circumflex, that is, if the pattern is
899     constrained to match only at the start of the subject, it is said to be an
900     "anchored" pattern. (There are also other constructs that can cause a pattern
901     to be anchored.)
902     </P>
903     <P>
904 nigel 75 A dollar character is an assertion that is true only if the current matching
905 nigel 63 point is at the end of the subject string, or immediately before a newline
906 nigel 91 at the end of the string (by default). Dollar need not be the last character of
907     the pattern if a number of alternatives are involved, but it should be the last
908     item in any branch in which it appears. Dollar has no special meaning in a
909     character class.
910 nigel 63 </P>
911     <P>
912     The meaning of dollar can be changed so that it matches only at the very end of
913     the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
914     does not affect the \Z assertion.
915     </P>
916     <P>
917     The meanings of the circumflex and dollar characters are changed if the
918 nigel 91 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
919     immediately after internal newlines as well as at the start of the subject
920     string. It does not match after a newline that ends the string. A dollar
921     matches before any newlines in the string, as well as at the very end, when
922     PCRE_MULTILINE is set. When newline is specified as the two-character
923     sequence CRLF, isolated CR and LF characters do not indicate newlines.
924 nigel 63 </P>
925     <P>
926 nigel 91 For example, the pattern /^abc$/ matches the subject string "def\nabc" (where
927     \n represents a newline) in multiline mode, but not otherwise. Consequently,
928     patterns that are anchored in single line mode because all branches start with
929     ^ are not anchored in multiline mode, and a match for circumflex is possible
930     when the <i>startoffset</i> argument of <b>pcre_exec()</b> is non-zero. The
931     PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
932     </P>
933     <P>
934 nigel 63 Note that the sequences \A, \Z, and \z can be used to match the start and
935     end of the subject in both modes, and if all branches of a pattern start with
936 nigel 91 \A it is always anchored, whether or not PCRE_MULTILINE is set.
937 ph10 518 <a name="fullstopdot"></a></P>
938     <br><a name="SEC6" href="#TOC1">FULL STOP (PERIOD, DOT) AND \N</a><br>
939 nigel 63 <P>
940     Outside a character class, a dot in the pattern matches any one character in
941 nigel 91 the subject string except (by default) a character that signifies the end of a
942 nigel 93 line. In UTF-8 mode, the matched character may be more than one byte long.
943 nigel 63 </P>
944 nigel 91 <P>
945 nigel 93 When a line ending is defined as a single character, dot never matches that
946     character; when the two-character sequence CRLF is used, dot does not match CR
947     if it is immediately followed by LF, but otherwise it matches all characters
948     (including isolated CRs and LFs). When any Unicode line endings are being
949     recognized, dot does not match CR or LF or any of the other line ending
950     characters.
951     </P>
952     <P>
953 nigel 91 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
954 nigel 93 option is set, a dot matches any one character, without exception. If the
955     two-character sequence CRLF is present in the subject string, it takes two dots
956     to match it.
957 nigel 91 </P>
958     <P>
959     The handling of dot is entirely independent of the handling of circumflex and
960     dollar, the only relationship being that they both involve newlines. Dot has no
961     special meaning in a character class.
962     </P>
963 ph10 518 <P>
964 ph10 579 The escape sequence \N behaves like a dot, except that it is not affected by
965 ph10 572 the PCRE_DOTALL option. In other words, it matches any character except one
966     that signifies the end of a line.
967 ph10 518 </P>
968 ph10 227 <br><a name="SEC7" href="#TOC1">MATCHING A SINGLE BYTE</a><br>
969 nigel 63 <P>
970     Outside a character class, the escape sequence \C matches any one byte, both
971 nigel 93 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
972     characters. The feature is provided in Perl in order to match individual bytes
973 ph10 572 in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes, the
974     rest of the string may start with a malformed UTF-8 character. For this reason,
975 nigel 93 the \C escape sequence is best avoided.
976 nigel 63 </P>
977     <P>
978 nigel 75 PCRE does not allow \C to appear in lookbehind assertions
979     <a href="#lookbehind">(described below),</a>
980     because in UTF-8 mode this would make it impossible to calculate the length of
981     the lookbehind.
982     <a name="characterclass"></a></P>
983 ph10 227 <br><a name="SEC8" href="#TOC1">SQUARE BRACKETS AND CHARACTER CLASSES</a><br>
984 nigel 63 <P>
985     An opening square bracket introduces a character class, terminated by a closing
986 ph10 461 square bracket. A closing square bracket on its own is not special by default.
987     However, if the PCRE_JAVASCRIPT_COMPAT option is set, a lone closing square
988     bracket causes a compile-time error. If a closing square bracket is required as
989     a member of the class, it should be the first data character in the class
990     (after an initial circumflex, if present) or escaped with a backslash.
991 nigel 63 </P>
992     <P>
993     A character class matches a single character in the subject. In UTF-8 mode, the
994 ph10 461 character may be more than one byte long. A matched character must be in the
995     set of characters defined by the class, unless the first character in the class
996 nigel 63 definition is a circumflex, in which case the subject character must not be in
997     the set defined by the class. If a circumflex is actually required as a member
998     of the class, ensure it is not the first character, or escape it with a
999     backslash.
1000     </P>
1001     <P>
1002     For example, the character class [aeiou] matches any lower case vowel, while
1003     [^aeiou] matches any character that is not a lower case vowel. Note that a
1004 nigel 75 circumflex is just a convenient notation for specifying the characters that
1005     are in the class by enumerating those that are not. A class that starts with a
1006 ph10 461 circumflex is not an assertion; it still consumes a character from the subject
1007 nigel 75 string, and therefore it fails if the current pointer is at the end of the
1008     string.
1009 nigel 63 </P>
1010     <P>
1011     In UTF-8 mode, characters with values greater than 255 can be included in a
1012     class as a literal string of bytes, or by using the \x{ escaping mechanism.
1013     </P>
1014     <P>
1015     When caseless matching is set, any letters in a class represent both their
1016     upper case and lower case versions, so for example, a caseless [aeiou] matches
1017     "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
1018 nigel 77 caseful version would. In UTF-8 mode, PCRE always understands the concept of
1019     case for characters whose values are less than 128, so caseless matching is
1020     always possible. For characters with higher values, the concept of case is
1021     supported if PCRE is compiled with Unicode property support, but not otherwise.
1022 ph10 461 If you want to use caseless matching in UTF8-mode for characters 128 and above,
1023     you must ensure that PCRE is compiled with Unicode property support as well as
1024     with UTF-8 support.
1025 nigel 63 </P>
1026     <P>
1027 nigel 93 Characters that might indicate line breaks are never treated in any special way
1028     when matching character classes, whatever line-ending sequence is in use, and
1029     whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
1030     such as [^a] always matches one of these characters.
1031 nigel 63 </P>
1032     <P>
1033     The minus (hyphen) character can be used to specify a range of characters in a
1034     character class. For example, [d-m] matches any letter between d and m,
1035     inclusive. If a minus character is required in a class, it must be escaped with
1036     a backslash or appear in a position where it cannot be interpreted as
1037     indicating a range, typically as the first or last character in the class.
1038     </P>
1039     <P>
1040     It is not possible to have the literal character "]" as the end character of a
1041     range. A pattern such as [W-]46] is interpreted as a class of two characters
1042     ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
1043     "-46]". However, if the "]" is escaped with a backslash it is interpreted as
1044 nigel 75 the end of range, so [W-\]46] is interpreted as a class containing a range
1045     followed by two other characters. The octal or hexadecimal representation of
1046     "]" can also be used to end a range.
1047 nigel 63 </P>
1048     <P>
1049     Ranges operate in the collating sequence of character values. They can also be
1050     used for characters specified numerically, for example [\000-\037]. In UTF-8
1051     mode, ranges can include characters whose values are greater than 255, for
1052     example [\x{100}-\x{2ff}].
1053     </P>
1054     <P>
1055     If a range that includes letters is used when caseless matching is set, it
1056     matches the letters in either case. For example, [W-c] is equivalent to
1057 nigel 75 [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
1058 ph10 142 tables for a French locale are in use, [\xc8-\xcb] matches accented E
1059 nigel 75 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
1060     characters with values greater than 128 only when it is compiled with Unicode
1061     property support.
1062 nigel 63 </P>
1063     <P>
1064 ph10 579 The character escape sequences \d, \D, \h, \H, \p, \P, \s, \S, \v,
1065     \V, \w, and \W may appear in a character class, and add the characters that
1066     they match to the class. For example, [\dABCDEF] matches any hexadecimal
1067     digit. In UTF-8 mode, the PCRE_UCP option affects the meanings of \d, \s, \w
1068     and their upper case partners, just as it does when they appear outside a
1069     character class, as described in the section entitled
1070     <a href="#genericchartypes">"Generic character types"</a>
1071     above. The escape sequence \b has a different meaning inside a character
1072     class; it matches the backspace character. The sequences \B, \N, \R, and \X
1073     are not special inside a character class. Like any other unrecognized escape
1074     sequences, they are treated as the literal characters "B", "N", "R", and "X" by
1075     default, but cause an error if the PCRE_EXTRA option is set.
1076     </P>
1077     <P>
1078     A circumflex can conveniently be used with the upper case character types to
1079 ph10 535 specify a more restricted set of characters than the matching lower case type.
1080 ph10 579 For example, the class [^\W_] matches any letter or digit, but not underscore,
1081     whereas [\w] includes underscore. A positive character class should be read as
1082     "something OR something OR ..." and a negative class as "NOT something AND NOT
1083     something AND NOT ...".
1084 nigel 63 </P>
1085     <P>
1086 nigel 75 The only metacharacters that are recognized in character classes are backslash,
1087     hyphen (only where it can be interpreted as specifying a range), circumflex
1088     (only at the start), opening square bracket (only when it can be interpreted as
1089     introducing a POSIX class name - see the next section), and the terminating
1090     closing square bracket. However, escaping other non-alphanumeric characters
1091     does no harm.
1092 nigel 63 </P>
1093 ph10 227 <br><a name="SEC9" href="#TOC1">POSIX CHARACTER CLASSES</a><br>
1094 nigel 63 <P>
1095 nigel 75 Perl supports the POSIX notation for character classes. This uses names
1096 nigel 63 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
1097     this notation. For example,
1098     <pre>
1099     [01[:alpha:]%]
1100 nigel 75 </pre>
1101 nigel 63 matches "0", "1", any alphabetic character, or "%". The supported class names
1102 ph10 535 are:
1103 nigel 63 <pre>
1104     alnum letters and digits
1105     alpha letters
1106     ascii character codes 0 - 127
1107     blank space or tab only
1108     cntrl control characters
1109     digit decimal digits (same as \d)
1110     graph printing characters, excluding space
1111     lower lower case letters
1112     print printing characters, including space
1113 ph10 535 punct printing characters, excluding letters and digits and space
1114 nigel 63 space white space (not quite the same as \s)
1115     upper upper case letters
1116     word "word" characters (same as \w)
1117     xdigit hexadecimal digits
1118 nigel 75 </pre>
1119 nigel 63 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
1120     space (32). Notice that this list includes the VT character (code 11). This
1121     makes "space" different to \s, which does not include VT (for Perl
1122     compatibility).
1123     </P>
1124     <P>
1125     The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
1126     5.8. Another Perl extension is negation, which is indicated by a ^ character
1127     after the colon. For example,
1128     <pre>
1129     [12[:^digit:]]
1130 nigel 75 </pre>
1131 nigel 63 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
1132     syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
1133     supported, and an error is given if they are encountered.
1134     </P>
1135     <P>
1136 ph10 535 By default, in UTF-8 mode, characters with values greater than 128 do not match
1137     any of the POSIX character classes. However, if the PCRE_UCP option is passed
1138     to <b>pcre_compile()</b>, some of the classes are changed so that Unicode
1139     character properties are used. This is achieved by replacing the POSIX classes
1140     by other sequences, as follows:
1141     <pre>
1142     [:alnum:] becomes \p{Xan}
1143     [:alpha:] becomes \p{L}
1144     [:blank:] becomes \h
1145     [:digit:] becomes \p{Nd}
1146     [:lower:] becomes \p{Ll}
1147     [:space:] becomes \p{Xps}
1148     [:upper:] becomes \p{Lu}
1149     [:word:] becomes \p{Xwd}
1150     </pre>
1151     Negated versions, such as [:^alpha:] use \P instead of \p. The other POSIX
1152     classes are unchanged, and match only characters with code points less than
1153     128.
1154 nigel 63 </P>
1155 ph10 227 <br><a name="SEC10" href="#TOC1">VERTICAL BAR</a><br>
1156 nigel 63 <P>
1157     Vertical bar characters are used to separate alternative patterns. For example,
1158     the pattern
1159     <pre>
1160     gilbert|sullivan
1161 nigel 75 </pre>
1162 nigel 63 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
1163 nigel 91 and an empty alternative is permitted (matching the empty string). The matching
1164     process tries each alternative in turn, from left to right, and the first one
1165     that succeeds is used. If the alternatives are within a subpattern
1166 nigel 75 <a href="#subpattern">(defined below),</a>
1167     "succeeds" means matching the rest of the main pattern as well as the
1168     alternative in the subpattern.
1169 nigel 63 </P>
1170 ph10 227 <br><a name="SEC11" href="#TOC1">INTERNAL OPTION SETTING</a><br>
1171 nigel 63 <P>
1172     The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1173 ph10 243 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1174     the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1175     The option letters are
1176 nigel 63 <pre>
1177     i for PCRE_CASELESS
1178     m for PCRE_MULTILINE
1179     s for PCRE_DOTALL
1180     x for PCRE_EXTENDED
1181 nigel 75 </pre>
1182 nigel 63 For example, (?im) sets caseless, multiline matching. It is also possible to
1183     unset these options by preceding the letter with a hyphen, and a combined
1184     setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1185     PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1186     permitted. If a letter appears both before and after the hyphen, the option is
1187     unset.
1188     </P>
1189     <P>
1190 ph10 243 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1191     changed in the same way as the Perl-compatible options by using the characters
1192     J, U and X respectively.
1193     </P>
1194     <P>
1195 ph10 416 When one of these option changes occurs at top level (that is, not inside
1196     subpattern parentheses), the change applies to the remainder of the pattern
1197     that follows. If the change is placed right at the start of a pattern, PCRE
1198     extracts it into the global options (and it will therefore show up in data
1199     extracted by the <b>pcre_fullinfo()</b> function).
1200 nigel 63 </P>
1201     <P>
1202 nigel 93 An option change within a subpattern (see below for a description of
1203 ph10 572 subpatterns) affects only that part of the subpattern that follows it, so
1204 nigel 63 <pre>
1205     (a(?i)b)c
1206 nigel 75 </pre>
1207 nigel 63 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1208     By this means, options can be made to have different settings in different
1209     parts of the pattern. Any changes made in one alternative do carry on
1210     into subsequent branches within the same subpattern. For example,
1211     <pre>
1212     (a(?i)b|c)
1213 nigel 75 </pre>
1214 nigel 63 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1215     branch is abandoned before the option setting. This is because the effects of
1216     option settings happen at compile time. There would be some very weird
1217     behaviour otherwise.
1218 ph10 254 </P>
1219     <P>
1220     <b>Note:</b> There are other PCRE-specific options that can be set by the
1221     application when the compile or match functions are called. In some cases the
1222 ph10 416 pattern can contain special leading sequences such as (*CRLF) to override what
1223     the application has set or what has been defaulted. Details are given in the
1224     section entitled
1225 ph10 254 <a href="#newlineseq">"Newline sequences"</a>
1226 ph10 535 above. There are also the (*UTF8) and (*UCP) leading sequences that can be used
1227     to set UTF-8 and Unicode property modes; they are equivalent to setting the
1228     PCRE_UTF8 and the PCRE_UCP options, respectively.
1229 nigel 75 <a name="subpattern"></a></P>
1230 ph10 227 <br><a name="SEC12" href="#TOC1">SUBPATTERNS</a><br>
1231 nigel 63 <P>
1232     Subpatterns are delimited by parentheses (round brackets), which can be nested.
1233 nigel 75 Turning part of a pattern into a subpattern does two things:
1234     <br>
1235     <br>
1236 nigel 63 1. It localizes a set of alternatives. For example, the pattern
1237     <pre>
1238     cat(aract|erpillar|)
1239 nigel 75 </pre>
1240 ph10 572 matches "cataract", "caterpillar", or "cat". Without the parentheses, it would
1241     match "cataract", "erpillar" or an empty string.
1242 nigel 75 <br>
1243     <br>
1244     2. It sets up the subpattern as a capturing subpattern. This means that, when
1245     the whole pattern matches, that portion of the subject string that matched the
1246     subpattern is passed back to the caller via the <i>ovector</i> argument of
1247 nigel 63 <b>pcre_exec()</b>. Opening parentheses are counted from left to right (starting
1248 ph10 572 from 1) to obtain numbers for the capturing subpatterns. For example, if the
1249     string "the red king" is matched against the pattern
1250 nigel 63 <pre>
1251     the ((red|white) (king|queen))
1252 nigel 75 </pre>
1253 nigel 63 the captured substrings are "red king", "red", and "king", and are numbered 1,
1254     2, and 3, respectively.
1255     </P>
1256     <P>
1257     The fact that plain parentheses fulfil two functions is not always helpful.
1258     There are often times when a grouping subpattern is required without a
1259     capturing requirement. If an opening parenthesis is followed by a question mark
1260     and a colon, the subpattern does not do any capturing, and is not counted when
1261     computing the number of any subsequent capturing subpatterns. For example, if
1262     the string "the white queen" is matched against the pattern
1263     <pre>
1264     the ((?:red|white) (king|queen))
1265 nigel 75 </pre>
1266 nigel 63 the captured substrings are "white queen" and "queen", and are numbered 1 and
1267 nigel 93 2. The maximum number of capturing subpatterns is 65535.
1268 nigel 63 </P>
1269     <P>
1270     As a convenient shorthand, if any option settings are required at the start of
1271     a non-capturing subpattern, the option letters may appear between the "?" and
1272     the ":". Thus the two patterns
1273     <pre>
1274     (?i:saturday|sunday)
1275     (?:(?i)saturday|sunday)
1276 nigel 75 </pre>
1277 nigel 63 match exactly the same set of strings. Because alternative branches are tried
1278     from left to right, and options are not reset until the end of the subpattern
1279     is reached, an option setting in one branch does affect subsequent branches, so
1280     the above patterns match "SUNDAY" as well as "Saturday".
1281 ph10 461 <a name="dupsubpatternnumber"></a></P>
1282 ph10 227 <br><a name="SEC13" href="#TOC1">DUPLICATE SUBPATTERN NUMBERS</a><br>
1283 nigel 63 <P>
1284 ph10 182 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1285     the same numbers for its capturing parentheses. Such a subpattern starts with
1286     (?| and is itself a non-capturing subpattern. For example, consider this
1287     pattern:
1288     <pre>
1289     (?|(Sat)ur|(Sun))day
1290     </pre>
1291     Because the two alternatives are inside a (?| group, both sets of capturing
1292     parentheses are numbered one. Thus, when the pattern matches, you can look
1293     at captured substring number one, whichever alternative matched. This construct
1294     is useful when you want to capture part, but not all, of one of a number of
1295     alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1296     number is reset at the start of each branch. The numbers of any capturing
1297 ph10 572 parentheses that follow the subpattern start after the highest number used in
1298     any branch. The following example is taken from the Perl documentation. The
1299     numbers underneath show in which buffer the captured content will be stored.
1300 ph10 182 <pre>
1301     # before ---------------branch-reset----------- after
1302     / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1303     # 1 2 2 3 2 3 4
1304     </pre>
1305 ph10 488 A back reference to a numbered subpattern uses the most recent value that is
1306     set for that number by any subpattern. The following pattern matches "abcabc"
1307     or "defdef":
1308 ph10 461 <pre>
1309     /(?|(abc)|(def))\1/
1310     </pre>
1311 ph10 733 In contrast, a subroutine call to a numbered subpattern always refers to the
1312     first one in the pattern with the given number. The following pattern matches
1313     "abcabc" or "defabc":
1314 ph10 461 <pre>
1315     /(?|(abc)|(def))(?1)/
1316     </pre>
1317     If a
1318     <a href="#conditions">condition test</a>
1319     for a subpattern's having matched refers to a non-unique number, the test is
1320     true if any of the subpatterns of that number have matched.
1321 ph10 182 </P>
1322     <P>
1323     An alternative approach to using this "branch reset" feature is to use
1324     duplicate named subpatterns, as described in the next section.
1325     </P>
1326 ph10 227 <br><a name="SEC14" href="#TOC1">NAMED SUBPATTERNS</a><br>
1327 ph10 182 <P>
1328 nigel 63 Identifying capturing parentheses by number is simple, but it can be very hard
1329     to keep track of the numbers in complicated regular expressions. Furthermore,
1330 nigel 75 if an expression is modified, the numbers may change. To help with this
1331 nigel 93 difficulty, PCRE supports the naming of subpatterns. This feature was not
1332     added to Perl until release 5.10. Python had the feature earlier, and PCRE
1333     introduced it at release 4.0, using the Python syntax. PCRE now supports both
1334 ph10 461 the Perl and the Python syntax. Perl allows identically numbered subpatterns to
1335     have different names, but PCRE does not.
1336 nigel 93 </P>
1337     <P>
1338     In PCRE, a subpattern can be named in one of three ways: (?&#60;name&#62;...) or
1339     (?'name'...) as in Perl, or (?P&#60;name&#62;...) as in Python. References to capturing
1340 nigel 91 parentheses from other parts of the pattern, such as
1341 ph10 488 <a href="#backreferences">back references,</a>
1342 nigel 91 <a href="#recursion">recursion,</a>
1343     and
1344     <a href="#conditions">conditions,</a>
1345     can be made by name as well as by number.
1346 nigel 63 </P>
1347     <P>
1348 nigel 91 Names consist of up to 32 alphanumeric characters and underscores. Named
1349 nigel 93 capturing parentheses are still allocated numbers as well as names, exactly as
1350     if the names were not present. The PCRE API provides function calls for
1351     extracting the name-to-number translation table from a compiled pattern. There
1352     is also a convenience function for extracting a captured substring by name.
1353 nigel 91 </P>
1354     <P>
1355     By default, a name must be unique within a pattern, but it is possible to relax
1356 ph10 461 this constraint by setting the PCRE_DUPNAMES option at compile time. (Duplicate
1357     names are also always permitted for subpatterns with the same number, set up as
1358     described in the previous section.) Duplicate names can be useful for patterns
1359     where only one instance of the named parentheses can match. Suppose you want to
1360     match the name of a weekday, either as a 3-letter abbreviation or as the full
1361     name, and in both cases you want to extract the abbreviation. This pattern
1362     (ignoring the line breaks) does the job:
1363 nigel 91 <pre>
1364 nigel 93 (?&#60;DN&#62;Mon|Fri|Sun)(?:day)?|
1365     (?&#60;DN&#62;Tue)(?:sday)?|
1366     (?&#60;DN&#62;Wed)(?:nesday)?|
1367     (?&#60;DN&#62;Thu)(?:rsday)?|
1368     (?&#60;DN&#62;Sat)(?:urday)?
1369 nigel 91 </pre>
1370     There are five capturing substrings, but only one is ever set after a match.
1371 ph10 182 (An alternative way of solving this problem is to use a "branch reset"
1372     subpattern, as described in the previous section.)
1373     </P>
1374     <P>
1375 nigel 91 The convenience function for extracting the data by name returns the substring
1376 nigel 93 for the first (and in this example, the only) subpattern of that name that
1377 ph10 461 matched. This saves searching to find which numbered subpattern it was.
1378     </P>
1379     <P>
1380 ph10 488 If you make a back reference to a non-unique named subpattern from elsewhere in
1381 ph10 461 the pattern, the one that corresponds to the first occurrence of the name is
1382     used. In the absence of duplicate numbers (see the previous section) this is
1383     the one with the lowest number. If you use a named reference in a condition
1384     test (see the
1385     <a href="#conditions">section about conditions</a>
1386     below), either to check whether a subpattern has matched, or to check for
1387     recursion, all subpatterns with the same name are tested. If the condition is
1388     true for any one of them, the overall condition is true. This is the same
1389     behaviour as testing by number. For further details of the interfaces for
1390     handling named subpatterns, see the
1391 nigel 63 <a href="pcreapi.html"><b>pcreapi</b></a>
1392     documentation.
1393     </P>
1394 ph10 392 <P>
1395     <b>Warning:</b> You cannot use different names to distinguish between two
1396 ph10 461 subpatterns with the same number because PCRE uses only the numbers when
1397     matching. For this reason, an error is given at compile time if different names
1398     are given to subpatterns with the same number. However, you can give the same
1399     name to subpatterns with the same number, even when PCRE_DUPNAMES is not set.
1400 ph10 392 </P>
1401 ph10 227 <br><a name="SEC15" href="#TOC1">REPETITION</a><br>
1402 nigel 63 <P>
1403     Repetition is specified by quantifiers, which can follow any of the following
1404     items:
1405     <pre>
1406     a literal data character
1407 nigel 93 the dot metacharacter
1408 nigel 63 the \C escape sequence
1409 nigel 75 the \X escape sequence (in UTF-8 mode with Unicode properties)
1410 nigel 93 the \R escape sequence
1411 ph10 572 an escape such as \d or \pL that matches a single character
1412 nigel 63 a character class
1413     a back reference (see next section)
1414 ph10 654 a parenthesized subpattern (including assertions)
1415 ph10 733 a subroutine call to a subpattern (recursive or otherwise)
1416 nigel 75 </pre>
1417 nigel 63 The general repetition quantifier specifies a minimum and maximum number of
1418     permitted matches, by giving the two numbers in curly brackets (braces),
1419     separated by a comma. The numbers must be less than 65536, and the first must
1420     be less than or equal to the second. For example:
1421     <pre>
1422     z{2,4}
1423 nigel 75 </pre>
1424 nigel 63 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1425     character. If the second number is omitted, but the comma is present, there is
1426     no upper limit; if the second number and the comma are both omitted, the
1427     quantifier specifies an exact number of required matches. Thus
1428     <pre>
1429     [aeiou]{3,}
1430 nigel 75 </pre>
1431 nigel 63 matches at least 3 successive vowels, but may match many more, while
1432     <pre>
1433     \d{8}
1434 nigel 75 </pre>
1435 nigel 63 matches exactly 8 digits. An opening curly bracket that appears in a position
1436     where a quantifier is not allowed, or one that does not match the syntax of a
1437     quantifier, is taken as a literal character. For example, {,6} is not a
1438     quantifier, but a literal string of four characters.
1439     </P>
1440     <P>
1441     In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1442     bytes. Thus, for example, \x{100}{2} matches two UTF-8 characters, each of
1443 nigel 75 which is represented by a two-byte sequence. Similarly, when Unicode property
1444     support is available, \X{3} matches three Unicode extended sequences, each of
1445     which may be several bytes long (and they may be of different lengths).
1446 nigel 63 </P>
1447     <P>
1448     The quantifier {0} is permitted, causing the expression to behave as if the
1449 ph10 345 previous item and the quantifier were not present. This may be useful for
1450     subpatterns that are referenced as
1451     <a href="#subpatternsassubroutines">subroutines</a>
1452 ph10 572 from elsewhere in the pattern (but see also the section entitled
1453     <a href="#subdefine">"Defining subpatterns for use by reference only"</a>
1454     below). Items other than subpatterns that have a {0} quantifier are omitted
1455     from the compiled pattern.
1456 nigel 63 </P>
1457     <P>
1458 nigel 93 For convenience, the three most common quantifiers have single-character
1459     abbreviations:
1460 nigel 63 <pre>
1461     * is equivalent to {0,}
1462     + is equivalent to {1,}
1463     ? is equivalent to {0,1}
1464 nigel 75 </pre>
1465 nigel 63 It is possible to construct infinite loops by following a subpattern that can
1466     match no characters with a quantifier that has no upper limit, for example:
1467     <pre>
1468     (a?)*
1469 nigel 75 </pre>
1470 nigel 63 Earlier versions of Perl and PCRE used to give an error at compile time for
1471     such patterns. However, because there are cases where this can be useful, such
1472     patterns are now accepted, but if any repetition of the subpattern does in fact
1473     match no characters, the loop is forcibly broken.
1474     </P>
1475     <P>
1476     By default, the quantifiers are "greedy", that is, they match as much as
1477     possible (up to the maximum number of permitted times), without causing the
1478     rest of the pattern to fail. The classic example of where this gives problems
1479 nigel 75 is in trying to match comments in C programs. These appear between /* and */
1480     and within the comment, individual * and / characters may appear. An attempt to
1481     match C comments by applying the pattern
1482 nigel 63 <pre>
1483     /\*.*\*/
1484 nigel 75 </pre>
1485 nigel 63 to the string
1486     <pre>
1487 nigel 75 /* first comment */ not comment /* second comment */
1488     </pre>
1489 nigel 63 fails, because it matches the entire string owing to the greediness of the .*
1490     item.
1491     </P>
1492     <P>
1493     However, if a quantifier is followed by a question mark, it ceases to be
1494     greedy, and instead matches the minimum number of times possible, so the
1495     pattern
1496     <pre>
1497     /\*.*?\*/
1498 nigel 75 </pre>
1499 nigel 63 does the right thing with the C comments. The meaning of the various
1500     quantifiers is not otherwise changed, just the preferred number of matches.
1501     Do not confuse this use of question mark with its use as a quantifier in its
1502     own right. Because it has two uses, it can sometimes appear doubled, as in
1503     <pre>
1504     \d??\d
1505 nigel 75 </pre>
1506 nigel 63 which matches one digit by preference, but can match two if that is the only
1507     way the rest of the pattern matches.
1508     </P>
1509     <P>
1510 nigel 93 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1511 nigel 63 the quantifiers are not greedy by default, but individual ones can be made
1512     greedy by following them with a question mark. In other words, it inverts the
1513     default behaviour.
1514     </P>
1515     <P>
1516     When a parenthesized subpattern is quantified with a minimum repeat count that
1517 nigel 75 is greater than 1 or with a limited maximum, more memory is required for the
1518 nigel 63 compiled pattern, in proportion to the size of the minimum or maximum.
1519     </P>
1520     <P>
1521     If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1522 nigel 93 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1523 nigel 63 implicitly anchored, because whatever follows will be tried against every
1524     character position in the subject string, so there is no point in retrying the
1525     overall match at any position after the first. PCRE normally treats such a
1526     pattern as though it were preceded by \A.
1527     </P>
1528     <P>
1529     In cases where it is known that the subject string contains no newlines, it is
1530     worth setting PCRE_DOTALL in order to obtain this optimization, or
1531     alternatively using ^ to indicate anchoring explicitly.
1532     </P>
1533     <P>
1534     However, there is one situation where the optimization cannot be used. When .*
1535 ph10 488 is inside capturing parentheses that are the subject of a back reference
1536 nigel 93 elsewhere in the pattern, a match at the start may fail where a later one
1537     succeeds. Consider, for example:
1538 nigel 63 <pre>
1539     (.*)abc\1
1540 nigel 75 </pre>
1541 nigel 63 If the subject is "xyz123abc123" the match point is the fourth character. For
1542     this reason, such a pattern is not implicitly anchored.
1543     </P>
1544     <P>
1545     When a capturing subpattern is repeated, the value captured is the substring
1546     that matched the final iteration. For example, after
1547     <pre>
1548     (tweedle[dume]{3}\s*)+
1549 nigel 75 </pre>
1550 nigel 63 has matched "tweedledum tweedledee" the value of the captured substring is
1551     "tweedledee". However, if there are nested capturing subpatterns, the
1552     corresponding captured values may have been set in previous iterations. For
1553     example, after
1554     <pre>
1555     /(a|(b))+/
1556 nigel 75 </pre>
1557 nigel 63 matches "aba" the value of the second captured substring is "b".
1558 nigel 75 <a name="atomicgroup"></a></P>
1559 ph10 227 <br><a name="SEC16" href="#TOC1">ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS</a><br>
1560 nigel 63 <P>
1561 nigel 93 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1562     repetition, failure of what follows normally causes the repeated item to be
1563     re-evaluated to see if a different number of repeats allows the rest of the
1564     pattern to match. Sometimes it is useful to prevent this, either to change the
1565     nature of the match, or to cause it fail earlier than it otherwise might, when
1566     the author of the pattern knows there is no point in carrying on.
1567 nigel 63 </P>
1568     <P>
1569     Consider, for example, the pattern \d+foo when applied to the subject line
1570     <pre>
1571     123456bar
1572 nigel 75 </pre>
1573 nigel 63 After matching all 6 digits and then failing to match "foo", the normal
1574     action of the matcher is to try again with only 5 digits matching the \d+
1575     item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1576     (a term taken from Jeffrey Friedl's book) provides the means for specifying
1577     that once a subpattern has matched, it is not to be re-evaluated in this way.
1578     </P>
1579     <P>
1580 nigel 93 If we use atomic grouping for the previous example, the matcher gives up
1581 nigel 63 immediately on failing to match "foo" the first time. The notation is a kind of
1582     special parenthesis, starting with (?&#62; as in this example:
1583     <pre>
1584 nigel 73 (?&#62;\d+)foo
1585 nigel 75 </pre>
1586 nigel 63 This kind of parenthesis "locks up" the part of the pattern it contains once
1587     it has matched, and a failure further into the pattern is prevented from
1588     backtracking into it. Backtracking past it to previous items, however, works as
1589     normal.
1590     </P>
1591     <P>
1592     An alternative description is that a subpattern of this type matches the string
1593     of characters that an identical standalone pattern would match, if anchored at
1594     the current point in the subject string.
1595     </P>
1596     <P>
1597     Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1598     the above example can be thought of as a maximizing repeat that must swallow
1599     everything it can. So, while both \d+ and \d+? are prepared to adjust the
1600     number of digits they match in order to make the rest of the pattern match,
1601     (?&#62;\d+) can only match an entire sequence of digits.
1602     </P>
1603     <P>
1604     Atomic groups in general can of course contain arbitrarily complicated
1605     subpatterns, and can be nested. However, when the subpattern for an atomic
1606     group is just a single repeated item, as in the example above, a simpler
1607     notation, called a "possessive quantifier" can be used. This consists of an
1608     additional + character following a quantifier. Using this notation, the
1609     previous example can be rewritten as
1610     <pre>
1611 nigel 75 \d++foo
1612     </pre>
1613 ph10 208 Note that a possessive quantifier can be used with an entire group, for
1614     example:
1615     <pre>
1616     (abc|xyz){2,3}+
1617     </pre>
1618 nigel 63 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1619     option is ignored. They are a convenient notation for the simpler forms of
1620 nigel 93 atomic group. However, there is no difference in the meaning of a possessive
1621     quantifier and the equivalent atomic group, though there may be a performance
1622     difference; possessive quantifiers should be slightly faster.
1623 nigel 63 </P>
1624     <P>
1625 nigel 93 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1626     Jeffrey Friedl originated the idea (and the name) in the first edition of his
1627     book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1628     package, and PCRE copied it from there. It ultimately found its way into Perl
1629     at release 5.10.
1630 nigel 63 </P>
1631     <P>
1632 nigel 93 PCRE has an optimization that automatically "possessifies" certain simple
1633     pattern constructs. For example, the sequence A+B is treated as A++B because
1634     there is no point in backtracking into a sequence of A's when B must follow.
1635     </P>
1636     <P>
1637 nigel 63 When a pattern contains an unlimited repeat inside a subpattern that can itself
1638     be repeated an unlimited number of times, the use of an atomic group is the
1639     only way to avoid some failing matches taking a very long time indeed. The
1640     pattern
1641     <pre>
1642     (\D+|&#60;\d+&#62;)*[!?]
1643 nigel 75 </pre>
1644 nigel 63 matches an unlimited number of substrings that either consist of non-digits, or
1645     digits enclosed in &#60;&#62;, followed by either ! or ?. When it matches, it runs
1646     quickly. However, if it is applied to
1647     <pre>
1648     aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1649 nigel 75 </pre>
1650 nigel 63 it takes a long time before reporting failure. This is because the string can
1651 nigel 75 be divided between the internal \D+ repeat and the external * repeat in a
1652     large number of ways, and all have to be tried. (The example uses [!?] rather
1653     than a single character at the end, because both PCRE and Perl have an
1654     optimization that allows for fast failure when a single character is used. They
1655     remember the last single character that is required for a match, and fail early
1656     if it is not present in the string.) If the pattern is changed so that it uses
1657     an atomic group, like this:
1658 nigel 63 <pre>
1659     ((?&#62;\D+)|&#60;\d+&#62;)*[!?]
1660 nigel 75 </pre>
1661 nigel 63 sequences of non-digits cannot be broken, and failure happens quickly.
1662 nigel 75 <a name="backreferences"></a></P>
1663 ph10 227 <br><a name="SEC17" href="#TOC1">BACK REFERENCES</a><br>
1664 nigel 63 <P>
1665     Outside a character class, a backslash followed by a digit greater than 0 (and
1666     possibly further digits) is a back reference to a capturing subpattern earlier
1667     (that is, to its left) in the pattern, provided there have been that many
1668     previous capturing left parentheses.
1669     </P>
1670     <P>
1671     However, if the decimal number following the backslash is less than 10, it is
1672     always taken as a back reference, and causes an error only if there are not
1673     that many capturing left parentheses in the entire pattern. In other words, the
1674     parentheses that are referenced need not be to the left of the reference for
1675 nigel 91 numbers less than 10. A "forward back reference" of this type can make sense
1676     when a repetition is involved and the subpattern to the right has participated
1677     in an earlier iteration.
1678     </P>
1679     <P>
1680 nigel 93 It is not possible to have a numerical "forward back reference" to a subpattern
1681     whose number is 10 or more using this syntax because a sequence such as \50 is
1682     interpreted as a character defined in octal. See the subsection entitled
1683 nigel 91 "Non-printing characters"
1684 nigel 75 <a href="#digitsafterbackslash">above</a>
1685 nigel 93 for further details of the handling of digits following a backslash. There is
1686     no such problem when named parentheses are used. A back reference to any
1687     subpattern is possible using named parentheses (see below).
1688 nigel 63 </P>
1689     <P>
1690 nigel 93 Another way of avoiding the ambiguity inherent in the use of digits following a
1691 ph10 572 backslash is to use the \g escape sequence. This escape must be followed by an
1692     unsigned number or a negative number, optionally enclosed in braces. These
1693     examples are all identical:
1694 nigel 93 <pre>
1695     (ring), \1
1696     (ring), \g1
1697     (ring), \g{1}
1698     </pre>
1699 ph10 208 An unsigned number specifies an absolute reference without the ambiguity that
1700     is present in the older syntax. It is also useful when literal digits follow
1701     the reference. A negative number is a relative reference. Consider this
1702     example:
1703 nigel 93 <pre>
1704     (abc(def)ghi)\g{-1}
1705     </pre>
1706     The sequence \g{-1} is a reference to the most recently started capturing
1707 ph10 579 subpattern before \g, that is, is it equivalent to \2 in this example.
1708     Similarly, \g{-2} would be equivalent to \1. The use of relative references
1709     can be helpful in long patterns, and also in patterns that are created by
1710     joining together fragments that contain references within themselves.
1711 nigel 93 </P>
1712     <P>
1713 nigel 63 A back reference matches whatever actually matched the capturing subpattern in
1714     the current subject string, rather than anything matching the subpattern
1715     itself (see
1716     <a href="#subpatternsassubroutines">"Subpatterns as subroutines"</a>
1717     below for a way of doing that). So the pattern
1718     <pre>
1719     (sens|respons)e and \1ibility
1720 nigel 75 </pre>
1721 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1722     "sense and responsibility". If caseful matching is in force at the time of the
1723     back reference, the case of letters is relevant. For example,
1724     <pre>
1725     ((?i)rah)\s+\1
1726 nigel 75 </pre>
1727 nigel 63 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1728     capturing subpattern is matched caselessly.
1729     </P>
1730     <P>
1731 ph10 172 There are several different ways of writing back references to named
1732     subpatterns. The .NET syntax \k{name} and the Perl syntax \k&#60;name&#62; or
1733     \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1734     back reference syntax, in which \g can be used for both numeric and named
1735     references, is also supported. We could rewrite the above example in any of
1736 nigel 93 the following ways:
1737 nigel 63 <pre>
1738 nigel 93 (?&#60;p1&#62;(?i)rah)\s+\k&#60;p1&#62;
1739 ph10 172 (?'p1'(?i)rah)\s+\k{p1}
1740 nigel 91 (?P&#60;p1&#62;(?i)rah)\s+(?P=p1)
1741 ph10 172 (?&#60;p1&#62;(?i)rah)\s+\g{p1}
1742 nigel 75 </pre>
1743 nigel 91 A subpattern that is referenced by name may appear in the pattern before or
1744     after the reference.
1745     </P>
1746     <P>
1747 nigel 63 There may be more than one back reference to the same subpattern. If a
1748     subpattern has not actually been used in a particular match, any back
1749 ph10 461 references to it always fail by default. For example, the pattern
1750 nigel 63 <pre>
1751     (a|(bc))\2
1752 nigel 75 </pre>
1753 ph10 461 always fails if it starts to match "a" rather than "bc". However, if the
1754     PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back reference to an
1755     unset value matches an empty string.
1756     </P>
1757     <P>
1758     Because there may be many capturing parentheses in a pattern, all digits
1759     following a backslash are taken as part of a potential back reference number.
1760     If the pattern continues with a digit character, some delimiter must be used to
1761     terminate the back reference. If the PCRE_EXTENDED option is set, this can be
1762     whitespace. Otherwise, the \g{ syntax or an empty comment (see
1763 nigel 75 <a href="#comments">"Comments"</a>
1764     below) can be used.
1765 nigel 63 </P>
1766 ph10 488 <br><b>
1767     Recursive back references
1768     </b><br>
1769 nigel 63 <P>
1770     A back reference that occurs inside the parentheses to which it refers fails
1771     when the subpattern is first used, so, for example, (a\1) never matches.
1772     However, such references can be useful inside repeated subpatterns. For
1773     example, the pattern
1774     <pre>
1775     (a|b\1)+
1776 nigel 75 </pre>
1777 nigel 63 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1778     the subpattern, the back reference matches the character string corresponding
1779     to the previous iteration. In order for this to work, the pattern must be such
1780     that the first iteration does not need to match the back reference. This can be
1781     done using alternation, as in the example above, or by a quantifier with a
1782     minimum of zero.
1783 ph10 488 </P>
1784     <P>
1785     Back references of this type cause the group that they reference to be treated
1786     as an
1787     <a href="#atomicgroup">atomic group.</a>
1788     Once the whole group has been matched, a subsequent matching failure cannot
1789     cause backtracking into the middle of the group.
1790 nigel 75 <a name="bigassertions"></a></P>
1791 ph10 227 <br><a name="SEC18" href="#TOC1">ASSERTIONS</a><br>
1792 nigel 63 <P>
1793     An assertion is a test on the characters following or preceding the current
1794     matching point that does not actually consume any characters. The simple
1795 nigel 75 assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are described
1796     <a href="#smallassertions">above.</a>
1797     </P>
1798     <P>
1799 nigel 63 More complicated assertions are coded as subpatterns. There are two kinds:
1800     those that look ahead of the current position in the subject string, and those
1801 nigel 75 that look behind it. An assertion subpattern is matched in the normal way,
1802     except that it does not cause the current matching position to be changed.
1803 nigel 63 </P>
1804     <P>
1805 ph10 654 Assertion subpatterns are not capturing subpatterns. If such an assertion
1806     contains capturing subpatterns within it, these are counted for the purposes of
1807     numbering the capturing subpatterns in the whole pattern. However, substring
1808     capturing is carried out only for positive assertions, because it does not make
1809     sense for negative assertions.
1810 nigel 63 </P>
1811 ph10 654 <P>
1812     For compatibility with Perl, assertion subpatterns may be repeated; though
1813     it makes no sense to assert the same thing several times, the side effect of
1814     capturing parentheses may occasionally be useful. In practice, there only three
1815     cases:
1816     <br>
1817     <br>
1818     (1) If the quantifier is {0}, the assertion is never obeyed during matching.
1819     However, it may contain internal capturing parenthesized groups that are called
1820     from elsewhere via the
1821     <a href="#subpatternsassubroutines">subroutine mechanism.</a>
1822     <br>
1823     <br>
1824     (2) If quantifier is {0,n} where n is greater than zero, it is treated as if it
1825     were {0,1}. At run time, the rest of the pattern match is tried with and
1826     without the assertion, the order depending on the greediness of the quantifier.
1827     <br>
1828     <br>
1829     (3) If the minimum repetition is greater than zero, the quantifier is ignored.
1830     The assertion is obeyed just once when encountered during matching.
1831     </P>
1832 nigel 75 <br><b>
1833     Lookahead assertions
1834     </b><br>
1835 nigel 63 <P>
1836 nigel 91 Lookahead assertions start with (?= for positive assertions and (?! for
1837     negative assertions. For example,
1838 nigel 63 <pre>
1839     \w+(?=;)
1840 nigel 75 </pre>
1841 nigel 63 matches a word followed by a semicolon, but does not include the semicolon in
1842     the match, and
1843     <pre>
1844     foo(?!bar)
1845 nigel 75 </pre>
1846 nigel 63 matches any occurrence of "foo" that is not followed by "bar". Note that the
1847     apparently similar pattern
1848     <pre>
1849     (?!foo)bar
1850 nigel 75 </pre>
1851 nigel 63 does not find an occurrence of "bar" that is preceded by something other than
1852     "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1853     (?!foo) is always true when the next three characters are "bar". A
1854 nigel 75 lookbehind assertion is needed to achieve the other effect.
1855 nigel 63 </P>
1856     <P>
1857     If you want to force a matching failure at some point in a pattern, the most
1858     convenient way to do it is with (?!) because an empty string always matches, so
1859     an assertion that requires there not to be an empty string must always fail.
1860 ph10 579 The backtracking control verb (*FAIL) or (*F) is a synonym for (?!).
1861 nigel 75 <a name="lookbehind"></a></P>
1862     <br><b>
1863     Lookbehind assertions
1864     </b><br>
1865 nigel 63 <P>
1866     Lookbehind assertions start with (?&#60;= for positive assertions and (?&#60;! for
1867     negative assertions. For example,
1868     <pre>
1869     (?&#60;!foo)bar
1870 nigel 75 </pre>
1871 nigel 63 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1872     a lookbehind assertion are restricted such that all the strings it matches must
1873 nigel 91 have a fixed length. However, if there are several top-level alternatives, they
1874     do not all have to have the same fixed length. Thus
1875 nigel 63 <pre>
1876     (?&#60;=bullock|donkey)
1877 nigel 75 </pre>
1878 nigel 63 is permitted, but
1879     <pre>
1880     (?&#60;!dogs?|cats?)
1881 nigel 75 </pre>
1882 nigel 63 causes an error at compile time. Branches that match different length strings
1883     are permitted only at the top level of a lookbehind assertion. This is an
1884 ph10 572 extension compared with Perl, which requires all branches to match the same
1885     length of string. An assertion such as
1886 nigel 63 <pre>
1887     (?&#60;=ab(c|de))
1888 nigel 75 </pre>
1889 nigel 63 is not permitted, because its single top-level branch can match two different
1890 ph10 454 lengths, but it is acceptable to PCRE if rewritten to use two top-level
1891     branches:
1892 nigel 63 <pre>
1893     (?&#60;=abc|abde)
1894 nigel 75 </pre>
1895 ph10 572 In some cases, the escape sequence \K
1896 ph10 172 <a href="#resetmatchstart">(see above)</a>
1897 ph10 461 can be used instead of a lookbehind assertion to get round the fixed-length
1898 ph10 454 restriction.
1899 ph10 172 </P>
1900     <P>
1901 nigel 63 The implementation of lookbehind assertions is, for each alternative, to
1902 nigel 93 temporarily move the current position back by the fixed length and then try to
1903 nigel 63 match. If there are insufficient characters before the current position, the
1904 nigel 93 assertion fails.
1905 nigel 63 </P>
1906     <P>
1907     PCRE does not allow the \C escape (which matches a single byte in UTF-8 mode)
1908     to appear in lookbehind assertions, because it makes it impossible to calculate
1909 nigel 93 the length of the lookbehind. The \X and \R escapes, which can match
1910     different numbers of bytes, are also not permitted.
1911 nigel 63 </P>
1912     <P>
1913 ph10 454 <a href="#subpatternsassubroutines">"Subroutine"</a>
1914     calls (see below) such as (?2) or (?&X) are permitted in lookbehinds, as long
1915 ph10 461 as the subpattern matches a fixed-length string.
1916 ph10 454 <a href="#recursion">Recursion,</a>
1917     however, is not supported.
1918     </P>
1919     <P>
1920 nigel 93 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1921 ph10 461 specify efficient matching of fixed-length strings at the end of subject
1922     strings. Consider a simple pattern such as
1923 nigel 63 <pre>
1924     abcd$
1925 nigel 75 </pre>
1926 nigel 63 when applied to a long string that does not match. Because matching proceeds
1927     from left to right, PCRE will look for each "a" in the subject and then see if
1928     what follows matches the rest of the pattern. If the pattern is specified as
1929     <pre>
1930     ^.*abcd$
1931 nigel 75 </pre>
1932 nigel 63 the initial .* matches the entire string at first, but when this fails (because
1933     there is no following "a"), it backtracks to match all but the last character,
1934     then all but the last two characters, and so on. Once again the search for "a"
1935     covers the entire string, from right to left, so we are no better off. However,
1936     if the pattern is written as
1937     <pre>
1938     ^.*+(?&#60;=abcd)
1939 nigel 75 </pre>
1940 nigel 93 there can be no backtracking for the .*+ item; it can match only the entire
1941 nigel 63 string. The subsequent lookbehind assertion does a single test on the last four
1942     characters. If it fails, the match fails immediately. For long strings, this
1943     approach makes a significant difference to the processing time.
1944     </P>
1945 nigel 75 <br><b>
1946     Using multiple assertions
1947     </b><br>
1948 nigel 63 <P>
1949     Several assertions (of any sort) may occur in succession. For example,
1950     <pre>
1951     (?&#60;=\d{3})(?&#60;!999)foo
1952 nigel 75 </pre>
1953 nigel 63 matches "foo" preceded by three digits that are not "999". Notice that each of
1954     the assertions is applied independently at the same point in the subject
1955     string. First there is a check that the previous three characters are all
1956     digits, and then there is a check that the same three characters are not "999".
1957     This pattern does <i>not</i> match "foo" preceded by six characters, the first
1958     of which are digits and the last three of which are not "999". For example, it
1959     doesn't match "123abcfoo". A pattern to do that is
1960     <pre>
1961     (?&#60;=\d{3}...)(?&#60;!999)foo
1962 nigel 75 </pre>
1963 nigel 63 This time the first assertion looks at the preceding six characters, checking
1964     that the first three are digits, and then the second assertion checks that the
1965     preceding three characters are not "999".
1966     </P>
1967     <P>
1968     Assertions can be nested in any combination. For example,
1969     <pre>
1970     (?&#60;=(?&#60;!foo)bar)baz
1971 nigel 75 </pre>
1972 nigel 63 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1973     preceded by "foo", while
1974     <pre>
1975     (?&#60;=\d{3}(?!999)...)foo
1976 nigel 75 </pre>
1977     is another pattern that matches "foo" preceded by three digits and any three
1978 nigel 63 characters that are not "999".
1979 nigel 91 <a name="conditions"></a></P>
1980 ph10 227 <br><a name="SEC19" href="#TOC1">CONDITIONAL SUBPATTERNS</a><br>
1981 nigel 63 <P>
1982     It is possible to cause the matching process to obey a subpattern
1983     conditionally or to choose between two alternative subpatterns, depending on
1984 ph10 461 the result of an assertion, or whether a specific capturing subpattern has
1985     already been matched. The two possible forms of conditional subpattern are:
1986 nigel 63 <pre>
1987     (?(condition)yes-pattern)
1988     (?(condition)yes-pattern|no-pattern)
1989 nigel 75 </pre>
1990 nigel 63 If the condition is satisfied, the yes-pattern is used; otherwise the
1991     no-pattern (if present) is used. If there are more than two alternatives in the
1992 ph10 567 subpattern, a compile-time error occurs. Each of the two alternatives may
1993 ph10 579 itself contain nested subpatterns of any form, including conditional
1994 ph10 567 subpatterns; the restriction to two alternatives applies only at the level of
1995 ph10 579 the condition. This pattern fragment is an example where the alternatives are
1996 ph10 567 complex:
1997     <pre>
1998     (?(1) (A|B|C) | (D | (?(2)E|F) | E) )
2000     </PRE>
2001 nigel 63 </P>
2002     <P>
2003 nigel 93 There are four kinds of condition: references to subpatterns, references to
2004     recursion, a pseudo-condition called DEFINE, and assertions.
2005 nigel 91 </P>
2006 nigel 93 <br><b>
2007     Checking for a used subpattern by number
2008     </b><br>
2009 nigel 91 <P>
2010 nigel 93 If the text between the parentheses consists of a sequence of digits, the
2011 ph10 461 condition is true if a capturing subpattern of that number has previously
2012     matched. If there is more than one capturing subpattern with the same number
2013     (see the earlier
2014     <a href="#recursion">section about duplicate subpattern numbers),</a>
2015 ph10 579 the condition is true if any of them have matched. An alternative notation is
2016 ph10 461 to precede the digits with a plus or minus sign. In this case, the subpattern
2017     number is relative rather than absolute. The most recently opened parentheses
2018 ph10 579 can be referenced by (?(-1), the next most recent by (?(-2), and so on. Inside
2019     loops it can also make sense to refer to subsequent groups. The next
2020     parentheses to be opened can be referenced as (?(+1), and so on. (The value
2021     zero in any of these forms is not used; it provokes a compile-time error.)
2022 nigel 93 </P>
2023     <P>
2024 nigel 91 Consider the following pattern, which contains non-significant white space to
2025     make it more readable (assume the PCRE_EXTENDED option) and to divide it into
2026     three parts for ease of discussion:
2027 nigel 63 <pre>
2028     ( \( )? [^()]+ (?(1) \) )
2029 nigel 75 </pre>
2030 nigel 63 The first part matches an optional opening parenthesis, and if that
2031     character is present, sets it as the first captured substring. The second part
2032     matches one or more characters that are not parentheses. The third part is a
2033 ph10 579 conditional subpattern that tests whether or not the first set of parentheses
2034     matched. If they did, that is, if subject started with an opening parenthesis,
2035 nigel 63 the condition is true, and so the yes-pattern is executed and a closing
2036     parenthesis is required. Otherwise, since no-pattern is not present, the
2037     subpattern matches nothing. In other words, this pattern matches a sequence of
2038 nigel 93 non-parentheses, optionally enclosed in parentheses.
2039     </P>
2040 ph10 172 <P>
2041     If you were embedding this pattern in a larger one, you could use a relative
2042     reference:
2043     <pre>
2044     ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
2045     </pre>
2046     This makes the fragment independent of the parentheses in the larger pattern.
2047     </P>
2048 nigel 93 <br><b>
2049     Checking for a used subpattern by name
2050     </b><br>
2051     <P>
2052     Perl uses the syntax (?(&#60;name&#62;)...) or (?('name')...) to test for a used
2053     subpattern by name. For compatibility with earlier versions of PCRE, which had
2054     this facility before Perl, the syntax (?(name)...) is also recognized. However,
2055     there is a possible ambiguity with this syntax, because subpattern names may
2056     consist entirely of digits. PCRE looks first for a named subpattern; if it
2057     cannot find one and the name consists entirely of digits, PCRE looks for a
2058     subpattern of that number, which must be greater than zero. Using subpattern
2059     names that consist entirely of digits is not recommended.
2060     </P>
2061     <P>
2062     Rewriting the above example to use a named subpattern gives this:
2063 nigel 91 <pre>
2064 nigel 93 (?&#60;OPEN&#62; \( )? [^()]+ (?(&#60;OPEN&#62;) \) )
2065 ph10 461 </pre>
2066     If the name used in a condition of this kind is a duplicate, the test is
2067     applied to all subpatterns of the same name, and is true if any one of them has
2068     matched.
2069 nigel 93 </P>
2070     <br><b>
2071     Checking for pattern recursion
2072     </b><br>
2073     <P>
2074     If the condition is the string (R), and there is no subpattern with the name R,
2075     the condition is true if a recursive call to the whole pattern or any
2076     subpattern has been made. If digits or a name preceded by ampersand follow the
2077     letter R, for example:
2078     <pre>
2079     (?(R3)...) or (?(R&name)...)
2080 nigel 91 </pre>
2081 ph10 461 the condition is true if the most recent recursion is into a subpattern whose
2082 nigel 93 number or name is given. This condition does not check the entire recursion
2083 ph10 461 stack. If the name used in a condition of this kind is a duplicate, the test is
2084     applied to all subpatterns of the same name, and is true if any one of them is
2085     the most recent recursion.
2086 nigel 63 </P>
2087     <P>
2088 ph10 461 At "top level", all these recursion test conditions are false.
2089     <a href="#recursion">The syntax for recursive patterns</a>
2090     is described below.
2091 ph10 572 <a name="subdefine"></a></P>
2092 nigel 93 <br><b>
2093     Defining subpatterns for use by reference only
2094     </b><br>
2095     <P>
2096     If the condition is the string (DEFINE), and there is no subpattern with the
2097     name DEFINE, the condition is always false. In this case, there may be only one
2098     alternative in the subpattern. It is always skipped if control reaches this
2099     point in the pattern; the idea of DEFINE is that it can be used to define
2100 ph10 733 subroutines that can be referenced from elsewhere. (The use of
2101     <a href="#subpatternsassubroutines">subroutines</a>
2102 ph10 579 is described below.) For example, a pattern to match an IPv4 address such as
2103     "" could be written like this (ignore whitespace and line
2104     breaks):
2105 nigel 93 <pre>
2106     (?(DEFINE) (?&#60;byte&#62; 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
2107     \b (?&byte) (\.(?&byte)){3} \b
2108     </pre>
2109     The first part of the pattern is a DEFINE group inside which a another group
2110     named "byte" is defined. This matches an individual component of an IPv4
2111     address (a number less than 256). When matching takes place, this part of the
2112 ph10 461 pattern is skipped because DEFINE acts like a false condition. The rest of the
2113     pattern uses references to the named group to match the four dot-separated
2114     components of an IPv4 address, insisting on a word boundary at each end.
2115 nigel 93 </P>
2116     <br><b>
2117     Assertion conditions
2118     </b><br>
2119     <P>
2120     If the condition is not in any of the above formats, it must be an assertion.
2121 nigel 63 This may be a positive or negative lookahead or lookbehind assertion. Consider
2122     this pattern, again containing non-significant white space, and with the two
2123     alternatives on the second line:
2124     <pre>
2125     (?(?=[^a-z]*[a-z])
2126     \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
2127 nigel 75 </pre>
2128 nigel 63 The condition is a positive lookahead assertion that matches an optional
2129     sequence of non-letters followed by a letter. In other words, it tests for the
2130     presence of at least one letter in the subject. If a letter is found, the
2131     subject is matched against the first alternative; otherwise it is matched
2132     against the second. This pattern matches strings in one of the two forms
2133     dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
2134 nigel 75 <a name="comments"></a></P>
2135 ph10 227 <br><a name="SEC20" href="#TOC1">COMMENTS</a><br>
2136 nigel 63 <P>
2137 ph10 579 There are two ways of including comments in patterns that are processed by
2138 ph10 567 PCRE. In both cases, the start of the comment must not be in a character class,
2139     nor in the middle of any other sequence of related characters such as (?: or a
2140     subpattern name or number. The characters that make up a comment play no part
2141     in the pattern matching.
2142 nigel 63 </P>
2143     <P>
2144 ph10 567 The sequence (?# marks the start of a comment that continues up to the next
2145     closing parenthesis. Nested parentheses are not permitted. If the PCRE_EXTENDED
2146     option is set, an unescaped # character also introduces a comment, which in
2147     this case continues to immediately after the next newline character or
2148     character sequence in the pattern. Which characters are interpreted as newlines
2149     is controlled by the options passed to <b>pcre_compile()</b> or by a special
2150     sequence at the start of the pattern, as described in the section entitled
2151 ph10 572 <a href="#newlines">"Newline conventions"</a>
2152     above. Note that the end of this type of comment is a literal newline sequence
2153     in the pattern; escape sequences that happen to represent a newline do not
2154     count. For example, consider this pattern when PCRE_EXTENDED is set, and the
2155     default newline convention is in force:
2156 ph10 567 <pre>
2157     abc #comment \n still comment
2158     </pre>
2159 ph10 579 On encountering the # character, <b>pcre_compile()</b> skips along, looking for
2160 ph10 567 a newline in the pattern. The sequence \n is still literal at this stage, so
2161     it does not terminate the comment. Only an actual character with the code value
2162     0x0a (the default newline) does so.
2163 nigel 91 <a name="recursion"></a></P>
2164 ph10 227 <br><a name="SEC21" href="#TOC1">RECURSIVE PATTERNS</a><br>
2165 nigel 63 <P>
2166     Consider the problem of matching a string in parentheses, allowing for
2167     unlimited nested parentheses. Without the use of recursion, the best that can
2168     be done is to use a pattern that matches up to some fixed depth of nesting. It
2169 nigel 93 is not possible to handle an arbitrary nesting depth.
2170     </P>
2171     <P>
2172     For some time, Perl has provided a facility that allows regular expressions to
2173     recurse (amongst other things). It does this by interpolating Perl code in the
2174     expression at run time, and the code can refer to the expression itself. A Perl
2175     pattern using code interpolation to solve the parentheses problem can be
2176     created like this:
2177 nigel 63 <pre>
2178     $re = qr{\( (?: (?&#62;[^()]+) | (?p{$re}) )* \)}x;
2179 nigel 75 </pre>
2180 nigel 63 The (?p{...}) item interpolates Perl code at run time, and in this case refers
2181 nigel 93 recursively to the pattern in which it appears.
2182 nigel 63 </P>
2183     <P>
2184 nigel 93 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
2185     supports special syntax for recursion of the entire pattern, and also for
2186     individual subpattern recursion. After its introduction in PCRE and Python,
2187 ph10 453 this kind of recursion was subsequently introduced into Perl at release 5.10.
2188 nigel 63 </P>
2189     <P>
2190 nigel 93 A special item that consists of (? followed by a number greater than zero and a
2191 ph10 733 closing parenthesis is a recursive subroutine call of the subpattern of the
2192     given number, provided that it occurs inside that subpattern. (If not, it is a
2193     <a href="#subpatternsassubroutines">non-recursive subroutine</a>
2194 nigel 93 call, which is described in the next section.) The special item (?R) or (?0) is
2195     a recursive call of the entire regular expression.
2196 nigel 87 </P>
2197     <P>
2198     This PCRE pattern solves the nested parentheses problem (assume the
2199     PCRE_EXTENDED option is set so that white space is ignored):
2200 nigel 63 <pre>
2201 ph10 461 \( ( [^()]++ | (?R) )* \)
2202 nigel 75 </pre>
2203 nigel 63 First it matches an opening parenthesis. Then it matches any number of
2204     substrings which can either be a sequence of non-parentheses, or a recursive
2205 nigel 87 match of the pattern itself (that is, a correctly parenthesized substring).
2206 ph10 461 Finally there is a closing parenthesis. Note the use of a possessive quantifier
2207     to avoid backtracking into sequences of non-parentheses.
2208 nigel 63 </P>
2209     <P>
2210     If this were part of a larger pattern, you would not want to recurse the entire
2211     pattern, so instead you could use this:
2212     <pre>
2213 ph10 461 ( \( ( [^()]++ | (?1) )* \) )
2214 nigel 75 </pre>
2215 nigel 63 We have put the pattern into parentheses, and caused the recursion to refer to
2216 ph10 172 them instead of the whole pattern.
2217     </P>
2218     <P>
2219     In a larger pattern, keeping track of parenthesis numbers can be tricky. This
2220 ph10 572 is made easier by the use of relative references. Instead of (?1) in the
2221     pattern above you can write (?-2) to refer to the second most recently opened
2222     parentheses preceding the recursion. In other words, a negative number counts
2223     capturing parentheses leftwards from the point at which it is encountered.
2224 ph10 172 </P>
2225     <P>
2226     It is also possible to refer to subsequently opened parentheses, by writing
2227     references such as (?+2). However, these cannot be recursive because the
2228     reference is not inside the parentheses that are referenced. They are always
2229 ph10 733 <a href="#subpatternsassubroutines">non-recursive subroutine</a>
2230 ph10 454 calls, as described in the next section.
2231 ph10 172 </P>
2232     <P>
2233     An alternative approach is to use named parentheses instead. The Perl syntax
2234     for this is (?&name); PCRE's earlier syntax (?P&#62;name) is also supported. We
2235     could rewrite the above example as follows:
2236 nigel 63 <pre>
2237 ph10 461 (?&#60;pn&#62; \( ( [^()]++ | (?&pn) )* \) )
2238 nigel 75 </pre>
2239 nigel 93 If there is more than one subpattern with the same name, the earliest one is
2240 ph10 172 used.
2241     </P>
2242     <P>
2243     This particular example pattern that we have been looking at contains nested
2244 ph10 461 unlimited repeats, and so the use of a possessive quantifier for matching
2245     strings of non-parentheses is important when applying the pattern to strings
2246     that do not match. For example, when this pattern is applied to
2247 nigel 63 <pre>
2248     (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
2249 nigel 75 </pre>
2250 ph10 461 it yields "no match" quickly. However, if a possessive quantifier is not used,
2251 nigel 63 the match runs for a very long time indeed because there are so many different
2252     ways the + and * repeats can carve up the subject, and all have to be tested
2253     before failure can be reported.
2254     </P>
2255     <P>
2256 ph10 469 At the end of a match, the values of capturing parentheses are those from
2257     the outermost level. If you want to obtain intermediate values, a callout
2258     function can be used (see below and the
2259 nigel 63 <a href="pcrecallout.html"><b>pcrecallout</b></a>
2260     documentation). If the pattern above is matched against
2261     <pre>
2262     (ab(cd)ef)
2263 nigel 75 </pre>
2264 ph10 469 the value for the inner capturing parentheses (numbered 2) is "ef", which is
2265     the last value taken on at the top level. If a capturing subpattern is not
2266 ph10 733 matched at the top level, its final captured value is unset, even if it was
2267     (temporarily) set at a deeper level during the matching process.
2268 nigel 63 </P>
2269     <P>
2270 ph10 469 If there are more than 15 capturing parentheses in a pattern, PCRE has to
2271     obtain extra memory to store data during a recursion, which it does by using
2272     <b>pcre_malloc</b>, freeing it via <b>pcre_free</b> afterwards. If no memory can
2273     be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
2274     </P>
2275     <P>
2276 nigel 63 Do not confuse the (?R) item with the condition (R), which tests for recursion.
2277     Consider this pattern, which matches text in angle brackets, allowing for
2278     arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2279     recursing), whereas any characters are permitted at the outer level.
2280     <pre>
2281     &#60; (?: (?(R) \d++ | [^&#60;&#62;]*+) | (?R)) * &#62;
2282 nigel 75 </pre>
2283 nigel 63 In this pattern, (?(R) is the start of a conditional subpattern, with two
2284     different alternatives for the recursive and non-recursive cases. The (?R) item
2285     is the actual recursive call.
2286 ph10 453 <a name="recursiondifference"></a></P>
2287     <br><b>
2288 ph10 733 Differences in recursion processing between PCRE and Perl
2289 ph10 453 </b><br>
2290     <P>
2291 ph10 733 Recursion processing in PCRE differs from Perl in two important ways. In PCRE
2292     (like Python, but unlike Perl), a recursive subpattern call is always treated
2293     as an atomic group. That is, once it has matched some of the subject string, it
2294     is never re-entered, even if it contains untried alternatives and there is a
2295     subsequent matching failure. This can be illustrated by the following pattern,
2296     which purports to match a palindromic string that contains an odd number of
2297     characters (for example, "a", "aba", "abcba", "abcdcba"):
2298 ph10 453 <pre>
2299     ^(.|(.)(?1)\2)$
2300     </pre>
2301 ph10 461 The idea is that it either matches a single character, or two identical
2302     characters surrounding a sub-palindrome. In Perl, this pattern works; in PCRE
2303 ph10 453 it does not if the pattern is longer than three characters. Consider the
2304     subject string "abcba":
2305     </P>
2306     <P>
2307 ph10 461 At the top level, the first character is matched, but as it is not at the end
2308 ph10 453 of the string, the first alternative fails; the second alternative is taken
2309     and the recursion kicks in. The recursive call to subpattern 1 successfully
2310     matches the next character ("b"). (Note that the beginning and end of line
2311     tests are not part of the recursion).
2312     </P>
2313     <P>
2314     Back at the top level, the next character ("c") is compared with what
2315 ph10 461 subpattern 2 matched, which was "a". This fails. Because the recursion is
2316 ph10 453 treated as an atomic group, there are now no backtracking points, and so the
2317     entire match fails. (Perl is able, at this point, to re-enter the recursion and
2318     try the second alternative.) However, if the pattern is written with the
2319     alternatives in the other order, things are different:
2320     <pre>
2321     ^((.)(?1)\2|.)$
2322     </pre>
2323 ph10 461 This time, the recursing alternative is tried first, and continues to recurse
2324     until it runs out of characters, at which point the recursion fails. But this
2325     time we do have another alternative to try at the higher level. That is the big
2326 ph10 453 difference: in the previous case the remaining alternative is at a deeper
2327     recursion level, which PCRE cannot use.
2328     </P>
2329     <P>
2330 ph10 579 To change the pattern so that it matches all palindromic strings, not just
2331     those with an odd number of characters, it is tempting to change the pattern to
2332     this:
2333 ph10 453 <pre>
2334     ^((.)(?1)\2|.?)$
2335     </pre>
2336 ph10 461 Again, this works in Perl, but not in PCRE, and for the same reason. When a
2337     deeper recursion has matched a single character, it cannot be entered again in
2338     order to match an empty string. The solution is to separate the two cases, and
2339 ph10 453 write out the odd and even cases as alternatives at the higher level:
2340     <pre>
2341     ^(?:((.)(?1)\2|)|((.)(?3)\4|.))
2342     </pre>
2343 ph10 461 If you want to match typical palindromic phrases, the pattern has to ignore all
2344 ph10 453 non-word characters, which can be done like this:
2345     <pre>
2346     ^\W*+(?:((.)\W*+(?1)\W*+\2|)|((.)\W*+(?3)\W*+\4|\W*+.\W*+))\W*+$
2347     </pre>
2348 ph10 461 If run with the PCRE_CASELESS option, this pattern matches phrases such as "A
2349     man, a plan, a canal: Panama!" and it works well in both PCRE and Perl. Note
2350     the use of the possessive quantifier *+ to avoid backtracking into sequences of
2351 ph10 453 non-word characters. Without this, PCRE takes a great deal longer (ten times or
2352     more) to match typical phrases, and Perl takes so long that you think it has
2353     gone into a loop.
2354 ph10 461 </P>
2355     <P>
2356     <b>WARNING</b>: The palindrome-matching patterns above work only if the subject
2357     string does not start with a palindrome that is shorter than the entire string.
2358     For example, although "abcba" is correctly matched, if the subject is "ababa",
2359     PCRE finds the palindrome "aba" at the start, then fails at top level because
2360     the end of the string does not follow. Once again, it cannot jump back into the
2361     recursion to try other alternatives, so the entire match fails.
2362 ph10 733 </P>
2363     <P>
2364     The second way in which PCRE and Perl differ in their recursion processing is
2365     in the handling of captured values. In Perl, when a subpattern is called
2366     recursively or as a subpattern (see the next section), it has no access to any
2367     values that were captured outside the recursion, whereas in PCRE these values
2368     can be referenced. Consider this pattern:
2369     <pre>
2370     ^(.)(\1|a(?2))
2371     </pre>
2372     In PCRE, this pattern matches "bab". The first capturing parentheses match "b",
2373     then in the second group, when the back reference \1 fails to match "b", the
2374     second alternative matches "a" and then recurses. In the recursion, \1 does
2375     now match "b" and so the whole match succeeds. In Perl, the pattern fails to
2376     match because inside the recursive call \1 cannot access the externally set
2377     value.
2378 nigel 75 <a name="subpatternsassubroutines"></a></P>
2379 ph10 227 <br><a name="SEC22" href="#TOC1">SUBPATTERNS AS SUBROUTINES</a><br>
2380 nigel 63 <P>
2381 ph10 733 If the syntax for a recursive subpattern call (either by number or by
2382 nigel 63 name) is used outside the parentheses to which it refers, it operates like a
2383 ph10 733 subroutine in a programming language. The called subpattern may be defined
2384 ph10 172 before or after the reference. A numbered reference can be absolute or
2385     relative, as in these examples:
2386 nigel 63 <pre>
2387 ph10 172 (...(absolute)...)...(?2)...
2388     (...(relative)...)...(?-1)...
2389     (...(?+1)...(relative)...
2390     </pre>
2391     An earlier example pointed out that the pattern
2392     <pre>
2393 nigel 63 (sens|respons)e and \1ibility
2394 nigel 75 </pre>
2395 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
2396     "sense and responsibility". If instead the pattern
2397     <pre>
2398     (sens|respons)e and (?1)ibility
2399 nigel 75 </pre>
2400 nigel 63 is used, it does match "sense and responsibility" as well as the other two
2401 nigel 93 strings. Another example is given in the discussion of DEFINE above.
2402 nigel 63 </P>
2403 nigel 87 <P>
2404 ph10 733 All subroutine calls, whether recursive or not, are always treated as atomic
2405     groups. That is, once a subroutine has matched some of the subject string, it
2406     is never re-entered, even if it contains untried alternatives and there is a
2407     subsequent matching failure. Any capturing parentheses that are set during the
2408     subroutine call revert to their previous values afterwards.
2409 nigel 87 </P>
2410 nigel 63 <P>
2411 ph10 733 Processing options such as case-independence are fixed when a subpattern is
2412     defined, so if it is used as a subroutine, such options cannot be changed for
2413     different calls. For example, consider this pattern:
2414 nigel 93 <pre>
2415 ph10 172 (abc)(?i:(?-1))
2416 nigel 93 </pre>
2417     It matches "abcabc". It does not match "abcABC" because the change of
2418     processing option does not affect the called subpattern.
2419 ph10 345 <a name="onigurumasubroutines"></a></P>
2420     <br><a name="SEC23" href="#TOC1">ONIGURUMA SUBROUTINE SYNTAX</a><br>
2421     <P>
2422     For compatibility with Oniguruma, the non-Perl syntax \g followed by a name or
2423     a number enclosed either in angle brackets or single quotes, is an alternative
2424     syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2425     are two of the examples used above, rewritten using this syntax:
2426     <pre>
2427     (?&#60;pn&#62; \( ( (?&#62;[^()]+) | \g&#60;pn&#62; )* \) )
2428     (sens|respons)e and \g'1'ibility
2429     </pre>
2430     PCRE supports an extension to Oniguruma: if a number is preceded by a
2431     plus or a minus sign it is taken as a relative reference. For example:
2432     <pre>
2433     (abc)(?i:\g&#60;-1&#62;)
2434     </pre>
2435     Note that \g{...} (Perl syntax) and \g&#60;...&#62; (Oniguruma syntax) are <i>not</i>
2436     synonymous. The former is a back reference; the latter is a subroutine call.
2437 nigel 93 </P>
2438 ph10 345 <br><a name="SEC24" href="#TOC1">CALLOUTS</a><br>
2439 nigel 93 <P>
2440 nigel 63 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2441     code to be obeyed in the middle of matching a regular expression. This makes it
2442     possible, amongst other things, to extract different substrings that match the
2443     same pair of parentheses when there is a repetition.
2444     </P>
2445     <P>
2446     PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2447     code. The feature is called "callout". The caller of PCRE provides an external
2448     function by putting its entry point in the global variable <i>pcre_callout</i>.
2449     By default, this variable contains NULL, which disables all calling out.
2450     </P>
2451     <P>
2452     Within a regular expression, (?C) indicates the points at which the external
2453     function is to be called. If you want to identify different callout points, you
2454     can put a number less than 256 after the letter C. The default value is zero.
2455     For example, this pattern has two callout points:
2456     <pre>
2457 ph10 155 (?C1)abc(?C2)def
2458 nigel 75 </pre>
2459     If the PCRE_AUTO_CALLOUT flag is passed to <b>pcre_compile()</b>, callouts are
2460     automatically installed before each item in the pattern. They are all numbered
2461     255.
2462 nigel 63 </P>
2463     <P>
2464     During matching, when PCRE reaches a callout point (and <i>pcre_callout</i> is
2465     set), the external function is called. It is provided with the number of the
2466 nigel 75 callout, the position in the pattern, and, optionally, one item of data
2467     originally supplied by the caller of <b>pcre_exec()</b>. The callout function
2468     may cause matching to proceed, to backtrack, or to fail altogether. A complete
2469     description of the interface to the callout function is given in the
2470 nigel 63 <a href="pcrecallout.html"><b>pcrecallout</b></a>
2471     documentation.
2472 ph10 512 <a name="backtrackcontrol"></a></P>
2473 ph10 345 <br><a name="SEC25" href="#TOC1">BACKTRACKING CONTROL</a><br>
2474 nigel 63 <P>
2475 ph10 211 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2476     are described in the Perl documentation as "experimental and subject to change
2477     or removal in a future version of Perl". It goes on to say: "Their usage in
2478     production code should be noted to avoid problems during upgrades." The same
2479     remarks apply to the PCRE features described in this section.
2480     </P>
2481     <P>
2482 ph10 345 Since these verbs are specifically related to backtracking, most of them can be
2483     used only when the pattern is to be matched using <b>pcre_exec()</b>, which uses
2484     a backtracking algorithm. With the exception of (*FAIL), which behaves like a
2485     failing negative assertion, they cause an error if encountered by
2486 ph10 211 <b>pcre_dfa_exec()</b>.
2487     </P>
2488     <P>
2489 ph10 733 If any of these verbs are used in an assertion or in a subpattern that is
2490     called as a subroutine (whether or not recursively), their effect is confined
2491     to that subpattern; it does not extend to the surrounding pattern, with one
2492     exception: a *MARK that is encountered in a positive assertion <i>is</i> passed
2493     back (compare capturing parentheses in assertions). Note that such subpatterns
2494     are processed as anchored at the point where they are tested. Note also that
2495     Perl's treatment of subroutines is different in some cases.
2496 ph10 453 </P>
2497     <P>
2498 ph10 211 The new verbs make use of what was previously invalid syntax: an opening
2499 ph10 512 parenthesis followed by an asterisk. They are generally of the form
2500     (*VERB) or (*VERB:NAME). Some may take either form, with differing behaviour,
2501 ph10 733 depending on whether or not an argument is present. A name is any sequence of
2502     characters that does not include a closing parenthesis. If the name is empty,
2503     that is, if the closing parenthesis immediately follows the colon, the effect
2504     is as if the colon were not there. Any number of these verbs may occur in a
2505     pattern.
2506 ph10 211 </P>
2507 ph10 512 <P>
2508     PCRE contains some optimizations that are used to speed up matching by running
2509     some checks at the start of each match attempt. For example, it may know the
2510     minimum length of matching subject, or that a particular character must be
2511     present. When one of these optimizations suppresses the running of a match, any
2512     included backtracking verbs will not, of course, be processed. You can suppress
2513     the start-of-match optimizations by setting the PCRE_NO_START_OPTIMIZE option
2514 ph10 579 when calling <b>pcre_compile()</b> or <b>pcre_exec()</b>, or by starting the
2515     pattern with (*NO_START_OPT).
2516 ph10 512 </P>
2517 ph10 211 <br><b>
2518     Verbs that act immediately
2519     </b><br>
2520     <P>
2521 ph10 512 The following verbs act as soon as they are encountered. They may not be
2522     followed by a name.
2523 ph10 211 <pre>
2524     (*ACCEPT)
2525     </pre>
2526     This verb causes the match to end successfully, skipping the remainder of the
2527 ph10 733 pattern. However, when it is inside a subpattern that is called as a
2528     subroutine, only that subpattern is ended successfully. Matching then continues
2529     at the outer level. If (*ACCEPT) is inside capturing parentheses, the data so
2530     far is captured. For example:
2531 ph10 211 <pre>
2532 ph10 453 A((?:A|B(*ACCEPT)|C)D)
2533 ph10 211 </pre>
2534 ph10 461 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is captured by
2535 ph10 453 the outer parentheses.
2536 ph10 211 <pre>
2537     (*FAIL) or (*F)
2538     </pre>
2539 ph10 733 This verb causes a matching failure, forcing backtracking to occur. It is
2540 ph10 211 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2541     probably useful only when combined with (?{}) or (??{}). Those are, of course,
2542     Perl features that are not present in PCRE. The nearest equivalent is the
2543     callout feature, as for example in this pattern:
2544     <pre>
2545     a+(?C)(*FAIL)
2546     </pre>
2547     A match with the string "aaaa" always fails, but the callout is taken before
2548     each backtrack happens (in this example, 10 times).
2549     </P>
2550     <br><b>
2551 ph10 512 Recording which path was taken
2552     </b><br>
2553     <P>
2554     There is one verb whose main purpose is to track how a match was arrived at,
2555     though it also has a secondary use in conjunction with advancing the match
2556     starting point (see (*SKIP) below).
2557     <pre>
2558     (*MARK:NAME) or (*:NAME)
2559     </pre>
2560     A name is always required with this verb. There may be as many instances of
2561     (*MARK) as you like in a pattern, and their names do not have to be unique.
2562     </P>
2563     <P>
2564     When a match succeeds, the name of the last-encountered (*MARK) is passed back
2565     to the caller via the <i>pcre_extra</i> data structure, as described in the
2566     <a href="pcreapi.html#extradata">section on <i>pcre_extra</i></a>
2567     in the
2568     <a href="pcreapi.html"><b>pcreapi</b></a>
2569     documentation. No data is returned for a partial match. Here is an example of
2570     <b>pcretest</b> output, where the /K modifier requests the retrieval and
2571     outputting of (*MARK) data:
2572     <pre>
2573     /X(*MARK:A)Y|X(*MARK:B)Z/K
2574     XY
2575     0: XY
2576     MK: A
2577     XZ
2578     0: XZ
2579     MK: B
2580     </pre>
2581     The (*MARK) name is tagged with "MK:" in this output, and in this example it
2582     indicates which of the two alternatives matched. This is a more efficient way
2583     of obtaining this information than putting each alternative in its own
2584     capturing parentheses.
2585     </P>
2586     <P>
2587 ph10 654 If (*MARK) is encountered in a positive assertion, its name is recorded and
2588     passed back if it is the last-encountered. This does not happen for negative
2589 ph10 733 assertions.
2590 ph10 654 </P>
2591     <P>
2592 ph10 512 A name may also be returned after a failed match if the final path through the
2593     pattern involves (*MARK). However, unless (*MARK) used in conjunction with
2594     (*COMMIT), this is unlikely to happen for an unanchored pattern because, as the
2595     starting point for matching is advanced, the final check is often with an empty
2596     string, causing a failure before (*MARK) is reached. For example:
2597     <pre>
2598     /X(*MARK:A)Y|X(*MARK:B)Z/K
2599     XP
2600     No match
2601     </pre>
2602     There are three potential starting points for this match (starting with X,
2603     starting with P, and with an empty string). If the pattern is anchored, the
2604     result is different:
2605     <pre>
2606     /^X(*MARK:A)Y|^X(*MARK:B)Z/K
2607     XP
2608     No match, mark = B
2609     </pre>
2610     PCRE's start-of-match optimizations can also interfere with this. For example,
2611     if, as a result of a call to <b>pcre_study()</b>, it knows the minimum
2612     subject length for a match, a shorter subject will not be scanned at all.
2613     </P>
2614     <P>
2615     Note that similar anomalies (though different in detail) exist in Perl, no
2616     doubt for the same reasons. The use of (*MARK) data after a failed match of an
2617     unanchored pattern is not recommended, unless (*COMMIT) is involved.
2618     </P>
2619     <br><b>
2620 ph10 211 Verbs that act after backtracking
2621     </b><br>
2622     <P>
2623     The following verbs do nothing when they are encountered. Matching continues
2624 ph10 512 with what follows, but if there is no subsequent match, causing a backtrack to
2625     the verb, a failure is forced. That is, backtracking cannot pass to the left of
2626     the verb. However, when one of these verbs appears inside an atomic group, its
2627     effect is confined to that group, because once the group has been matched,
2628     there is never any backtracking into it. In this situation, backtracking can
2629     "jump back" to the left of the entire atomic group. (Remember also, as stated
2630     above, that this localization also applies in subroutine calls and assertions.)
2631     </P>
2632     <P>
2633     These verbs differ in exactly what kind of failure occurs when backtracking
2634     reaches them.
2635 ph10 211 <pre>
2636     (*COMMIT)
2637     </pre>
2638 ph10 512 This verb, which may not be followed by a name, causes the whole match to fail
2639     outright if the rest of the pattern does not match. Even if the pattern is
2640     unanchored, no further attempts to find a match by advancing the starting point
2641     take place. Once (*COMMIT) has been passed, <b>pcre_exec()</b> is committed to
2642     finding a match at the current starting point, or not at all. For example:
2643 ph10 211 <pre>
2644     a+(*COMMIT)b
2645     </pre>
2646     This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2647 ph10 512 dynamic anchor, or "I've started, so I must finish." The name of the most
2648     recently passed (*MARK) in the path is passed back when (*COMMIT) forces a
2649     match failure.
2650     </P>
2651     <P>
2652     Note that (*COMMIT) at the start of a pattern is not the same as an anchor,
2653     unless PCRE's start-of-match optimizations are turned off, as shown in this
2654     <b>pcretest</b> example:
2655 ph10 211 <pre>
2656 ph10 512 /(*COMMIT)abc/
2657     xyzabc
2658     0: abc
2659     xyzabc\Y
2660     No match
2661 ph10 211 </pre>
2662 ph10 512 PCRE knows that any match must start with "a", so the optimization skips along
2663     the subject to "a" before running the first match attempt, which succeeds. When
2664     the optimization is disabled by the \Y escape in the second subject, the match
2665     starts at "x" and so the (*COMMIT) causes it to fail without trying any other
2666     starting points.
2667 ph10 211 <pre>
2668 ph10 512 (*PRUNE) or (*PRUNE:NAME)
2669     </pre>
2670     This verb causes the match to fail at the current starting position in the
2671     subject if the rest of the pattern does not match. If the pattern is
2672     unanchored, the normal "bumpalong" advance to the next starting character then
2673     happens. Backtracking can occur as usual to the left of (*PRUNE), before it is
2674     reached, or when matching to the right of (*PRUNE), but if there is no match to
2675     the right, backtracking cannot cross (*PRUNE). In simple cases, the use of
2676     (*PRUNE) is just an alternative to an atomic group or possessive quantifier,
2677     but there are some uses of (*PRUNE) that cannot be expressed in any other way.
2678     The behaviour of (*PRUNE:NAME) is the same as (*MARK:NAME)(*PRUNE) when the
2679     match fails completely; the name is passed back if this is the final attempt.
2680     (*PRUNE:NAME) does not pass back a name if the match succeeds. In an anchored
2681     pattern (*PRUNE) has the same effect as (*COMMIT).
2682     <pre>
2683 ph10 211 (*SKIP)
2684     </pre>
2685 ph10 512 This verb, when given without a name, is like (*PRUNE), except that if the
2686     pattern is unanchored, the "bumpalong" advance is not to the next character,
2687     but to the position in the subject where (*SKIP) was encountered. (*SKIP)
2688     signifies that whatever text was matched leading up to it cannot be part of a
2689     successful match. Consider:
2690 ph10 211 <pre>
2691     a+(*SKIP)b
2692     </pre>
2693     If the subject is "aaaac...", after the first match attempt fails (starting at
2694     the first character in the string), the starting point skips on to start the
2695     next attempt at "c". Note that a possessive quantifer does not have the same
2696 ph10 461 effect as this example; although it would suppress backtracking during the
2697 ph10 211 first match attempt, the second attempt would start at the second character
2698     instead of skipping on to "c".
2699     <pre>
2700 ph10 512 (*SKIP:NAME)
2701 ph10 211 </pre>
2702 ph10 512 When (*SKIP) has an associated name, its behaviour is modified. If the
2703     following pattern fails to match, the previous path through the pattern is
2704     searched for the most recent (*MARK) that has the same name. If one is found,
2705     the "bumpalong" advance is to the subject position that corresponds to that
2706     (*MARK) instead of to where (*SKIP) was encountered. If no (*MARK) with a
2707 ph10 733 matching name is found, normal "bumpalong" of one character happens (that is,
2708     the (*SKIP) is ignored).
2709 ph10 512 <pre>
2710     (*THEN) or (*THEN:NAME)
2711     </pre>
2712 ph10 733 This verb causes a skip to the next innermost alternative if the rest of the
2713     pattern does not match. That is, it cancels pending backtracking, but only
2714     within the current alternative. Its name comes from the observation that it can
2715     be used for a pattern-based if-then-else block:
2716 ph10 211 <pre>
2717     ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
2718     </pre>
2719     If the COND1 pattern matches, FOO is tried (and possibly further items after
2720 ph10 733 the end of the group if FOO succeeds); on failure, the matcher skips to the
2721 ph10 512 second alternative and tries COND2, without backtracking into COND1. The
2722     behaviour of (*THEN:NAME) is exactly the same as (*MARK:NAME)(*THEN) if the
2723 ph10 733 overall match fails. If (*THEN) is not inside an alternation, it acts like
2724     (*PRUNE).
2725 ph10 211 </P>
2726 ph10 567 <P>
2727 ph10 733 Note that a subpattern that does not contain a | character is just a part of
2728     the enclosing alternative; it is not a nested alternation with only one
2729     alternative. The effect of (*THEN) extends beyond such a subpattern to the
2730     enclosing alternative. Consider this pattern, where A, B, etc. are complex
2731     pattern fragments that do not contain any | characters at this level:
2732     <pre>
2733     A (B(*THEN)C) | D
2734     </pre>
2735     If A and B are matched, but there is a failure in C, matching does not
2736     backtrack into A; instead it moves to the next alternative, that is, D.
2737     However, if the subpattern containing (*THEN) is given an alternative, it
2738     behaves differently:
2739     <pre>
2740     A (B(*THEN)C | (*FAIL)) | D
2741     </pre>
2742     The effect of (*THEN) is now confined to the inner subpattern. After a failure
2743     in C, matching moves to (*FAIL), which causes the whole subpattern to fail
2744     because there are no more alternatives to try. In this case, matching does now
2745     backtrack into A.
2746 ph10 567 </P>
2747     <P>
2748 ph10 733 Note also that a conditional subpattern is not considered as having two
2749     alternatives, because only one is ever used. In other words, the | character in
2750     a conditional subpattern has a different meaning. Ignoring white space,
2751     consider:
2752 ph10 567 <pre>
2753 ph10 733 ^.*? (?(?=a) a | b(*THEN)c )
2754     </pre>
2755     If the subject is "ba", this pattern does not match. Because .*? is ungreedy,
2756     it initially matches zero characters. The condition (?=a) then fails, the
2757     character "b" is matched, but "c" is not. At this point, matching does not
2758     backtrack to .*? as might perhaps be expected from the presence of the |
2759     character. The conditional subpattern is part of the single alternative that
2760     comprises the whole pattern, and so the match fails. (If there was a backtrack
2761     into .*?, allowing it to match "b", the match would succeed.)
2762     </P>
2763     <P>
2764     The verbs just described provide four different "strengths" of control when
2765     subsequent matching fails. (*THEN) is the weakest, carrying on the match at the
2766     next alternative. (*PRUNE) comes next, failing the match at the current
2767     starting position, but allowing an advance to the next character (for an
2768     unanchored pattern). (*SKIP) is similar, except that the advance may be more
2769     than one character. (*COMMIT) is the strongest, causing the entire match to
2770     fail.
2771     </P>
2772     <P>
2773     If more than one such verb is present in a pattern, the "strongest" one wins.
2774     For example, consider this pattern, where A, B, etc. are complex pattern
2775     fragments:
2776     <pre>
2777 ph10 567 (A(*COMMIT)B(*THEN)C|D)
2778     </pre>
2779     Once A has matched, PCRE is committed to this match, at the current starting
2780     position. If subsequently B matches, but C does not, the normal (*THEN) action
2781 ph10 733 of trying the next alternative (that is, D) does not happen because (*COMMIT)
2782 ph10 567 overrides.
2783     </P>
2784 ph10 345 <br><a name="SEC26" href="#TOC1">SEE ALSO</a><br>
2785 ph10 211 <P>
2786 ph10 461 <b>pcreapi</b>(3), <b>pcrecallout</b>(3), <b>pcrematching</b>(3),
2787     <b>pcresyntax</b>(3), <b>pcre</b>(3).
2788 nigel 93 </P>
2789 ph10 345 <br><a name="SEC27" href="#TOC1">AUTHOR</a><br>
2790 nigel 93 <P>
2791 ph10 99 Philip Hazel
2792 nigel 63 <br>
2793 ph10 99 University Computing Service
2794     <br>
2795     Cambridge CB2 3QH, England.
2796     <br>
2797     </P>
2798 ph10 345 <br><a name="SEC28" href="#TOC1">REVISION</a><br>
2799 ph10 99 <P>
2800 ph10 733 Last updated: 09 October 2011
2801 ph10 99 <br>
2802 ph10 654 Copyright &copy; 1997-2011 University of Cambridge.
2803 ph10 99 <br>
2804 nigel 75 <p>
2805     Return to the <a href="index.html">PCRE index page</a>.
2806     </p>


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

ViewVC Help
Powered by ViewVC 1.1.12