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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 142 - (hide annotations) (download) (as text)
Fri Mar 30 15:55:18 2007 UTC (7 years ago) by ph10
File MIME type: text/html
File size: 77989 byte(s)
Trailing spaces.

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 nigel 93 <li><a name="TOC2" href="#SEC2">CHARACTERS AND METACHARACTERS</a>
18     <li><a name="TOC3" href="#SEC3">BACKSLASH</a>
19     <li><a name="TOC4" href="#SEC4">CIRCUMFLEX AND DOLLAR</a>
20     <li><a name="TOC5" href="#SEC5">FULL STOP (PERIOD, DOT)</a>
21     <li><a name="TOC6" href="#SEC6">MATCHING A SINGLE BYTE</a>
22     <li><a name="TOC7" href="#SEC7">SQUARE BRACKETS AND CHARACTER CLASSES</a>
23     <li><a name="TOC8" href="#SEC8">POSIX CHARACTER CLASSES</a>
24     <li><a name="TOC9" href="#SEC9">VERTICAL BAR</a>
25     <li><a name="TOC10" href="#SEC10">INTERNAL OPTION SETTING</a>
26     <li><a name="TOC11" href="#SEC11">SUBPATTERNS</a>
27     <li><a name="TOC12" href="#SEC12">NAMED SUBPATTERNS</a>
28     <li><a name="TOC13" href="#SEC13">REPETITION</a>
29     <li><a name="TOC14" href="#SEC14">ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS</a>
30     <li><a name="TOC15" href="#SEC15">BACK REFERENCES</a>
31     <li><a name="TOC16" href="#SEC16">ASSERTIONS</a>
32     <li><a name="TOC17" href="#SEC17">CONDITIONAL SUBPATTERNS</a>
33     <li><a name="TOC18" href="#SEC18">COMMENTS</a>
34     <li><a name="TOC19" href="#SEC19">RECURSIVE PATTERNS</a>
35     <li><a name="TOC20" href="#SEC20">SUBPATTERNS AS SUBROUTINES</a>
36     <li><a name="TOC21" href="#SEC21">CALLOUTS</a>
37     <li><a name="TOC22" href="#SEC22">SEE ALSO</a>
38 ph10 99 <li><a name="TOC23" href="#SEC23">AUTHOR</a>
39     <li><a name="TOC24" href="#SEC24">REVISION</a>
40 nigel 63 </ul>
41     <br><a name="SEC1" href="#TOC1">PCRE REGULAR EXPRESSION DETAILS</a><br>
42     <P>
43     The syntax and semantics of the regular expressions supported by PCRE are
44     described below. Regular expressions are also described in the Perl
45 nigel 75 documentation and in a number of books, some of which have copious examples.
46     Jeffrey Friedl's "Mastering Regular Expressions", published by O'Reilly, covers
47     regular expressions in great detail. This description of PCRE's regular
48     expressions is intended as reference material.
49 nigel 63 </P>
50     <P>
51 nigel 75 The original operation of PCRE was on strings of one-byte characters. However,
52     there is now also support for UTF-8 character strings. To use this, you must
53     build PCRE to include UTF-8 support, and then call <b>pcre_compile()</b> with
54     the PCRE_UTF8 option. How this affects pattern matching is mentioned in several
55     places below. There is also a summary of UTF-8 features in the
56 nigel 63 <a href="pcre.html#utf8support">section on UTF-8 support</a>
57     in the main
58     <a href="pcre.html"><b>pcre</b></a>
59     page.
60     </P>
61     <P>
62 nigel 77 The remainder of this document discusses the patterns that are supported by
63     PCRE when its main matching function, <b>pcre_exec()</b>, is used.
64     From release 6.0, PCRE offers a second matching function,
65     <b>pcre_dfa_exec()</b>, which matches using a different algorithm that is not
66     Perl-compatible. The advantages and disadvantages of the alternative function,
67     and how it differs from the normal function, are discussed in the
68     <a href="pcrematching.html"><b>pcrematching</b></a>
69     page.
70     </P>
71 nigel 93 <br><a name="SEC2" href="#TOC1">CHARACTERS AND METACHARACTERS</a><br>
72 nigel 77 <P>
73 nigel 63 A regular expression is a pattern that is matched against a subject string from
74     left to right. Most characters stand for themselves in a pattern, and match the
75     corresponding characters in the subject. As a trivial example, the pattern
76     <pre>
77     The quick brown fox
78 nigel 75 </pre>
79 nigel 77 matches a portion of a subject string that is identical to itself. When
80     caseless matching is specified (the PCRE_CASELESS option), letters are matched
81     independently of case. In UTF-8 mode, PCRE always understands the concept of
82     case for characters whose values are less than 128, so caseless matching is
83     always possible. For characters with higher values, the concept of case is
84     supported if PCRE is compiled with Unicode property support, but not otherwise.
85     If you want to use caseless matching for characters 128 and above, you must
86     ensure that PCRE is compiled with Unicode property support as well as with
87     UTF-8 support.
88     </P>
89     <P>
90     The power of regular expressions comes from the ability to include alternatives
91     and repetitions in the pattern. These are encoded in the pattern by the use of
92 nigel 75 <i>metacharacters</i>, which do not stand for themselves but instead are
93 nigel 63 interpreted in some special way.
94     </P>
95     <P>
96 nigel 75 There are two different sets of metacharacters: those that are recognized
97 nigel 63 anywhere in the pattern except within square brackets, and those that are
98 nigel 93 recognized within square brackets. Outside square brackets, the metacharacters
99     are as follows:
100 nigel 63 <pre>
101     \ general escape character with several uses
102     ^ assert start of string (or line, in multiline mode)
103     $ assert end of string (or line, in multiline mode)
104     . match any character except newline (by default)
105     [ start character class definition
106     | start of alternative branch
107     ( start subpattern
108     ) end subpattern
109     ? extends the meaning of (
110     also 0 or 1 quantifier
111     also quantifier minimizer
112     * 0 or more quantifier
113     + 1 or more quantifier
114     also "possessive quantifier"
115     { start min/max quantifier
116 nigel 75 </pre>
117 nigel 63 Part of a pattern that is in square brackets is called a "character class". In
118 nigel 75 a character class the only metacharacters are:
119 nigel 63 <pre>
120     \ general escape character
121     ^ negate the class, but only if the first character
122     - indicates character range
123 nigel 75 [ POSIX character class (only if followed by POSIX syntax)
124 nigel 63 ] terminates the character class
125 nigel 75 </pre>
126     The following sections describe the use of each of the metacharacters.
127 nigel 63 </P>
128 nigel 93 <br><a name="SEC3" href="#TOC1">BACKSLASH</a><br>
129 nigel 63 <P>
130     The backslash character has several uses. Firstly, if it is followed by a
131 nigel 91 non-alphanumeric character, it takes away any special meaning that character
132     may have. This use of backslash as an escape character applies both inside and
133 nigel 63 outside character classes.
134     </P>
135     <P>
136     For example, if you want to match a * character, you write \* in the pattern.
137     This escaping action applies whether or not the following character would
138 nigel 75 otherwise be interpreted as a metacharacter, so it is always safe to precede a
139     non-alphanumeric with backslash to specify that it stands for itself. In
140 nigel 63 particular, if you want to match a backslash, you write \\.
141     </P>
142     <P>
143     If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
144     pattern (other than in a character class) and characters between a # outside
145 nigel 91 a character class and the next newline are ignored. An escaping backslash can
146     be used to include a whitespace or # character as part of the pattern.
147 nigel 63 </P>
148     <P>
149     If you want to remove the special meaning from a sequence of characters, you
150     can do so by putting them between \Q and \E. This is different from Perl in
151     that $ and @ are handled as literals in \Q...\E sequences in PCRE, whereas in
152     Perl, $ and @ cause variable interpolation. Note the following examples:
153     <pre>
154     Pattern PCRE matches Perl matches
155 nigel 75
156     \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
157 nigel 63 \Qabc\$xyz\E abc\$xyz abc\$xyz
158     \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
159 nigel 75 </pre>
160 nigel 63 The \Q...\E sequence is recognized both inside and outside character classes.
161 nigel 75 <a name="digitsafterbackslash"></a></P>
162     <br><b>
163     Non-printing characters
164     </b><br>
165 nigel 63 <P>
166     A second use of backslash provides a way of encoding non-printing characters
167     in patterns in a visible manner. There is no restriction on the appearance of
168     non-printing characters, apart from the binary zero that terminates a pattern,
169     but when a pattern is being prepared by text editing, it is usually easier to
170     use one of the following escape sequences than the binary character it
171     represents:
172     <pre>
173     \a alarm, that is, the BEL character (hex 07)
174     \cx "control-x", where x is any character
175     \e escape (hex 1B)
176     \f formfeed (hex 0C)
177     \n newline (hex 0A)
178     \r carriage return (hex 0D)
179     \t tab (hex 09)
180     \ddd character with octal code ddd, or backreference
181     \xhh character with hex code hh
182 nigel 87 \x{hhh..} character with hex code hhh..
183 nigel 75 </pre>
184 nigel 63 The precise effect of \cx is as follows: if x is a lower case letter, it
185     is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
186     Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c; becomes hex
187     7B.
188     </P>
189     <P>
190     After \x, from zero to two hexadecimal digits are read (letters can be in
191 nigel 87 upper or lower case). Any number of hexadecimal digits may appear between \x{
192     and }, but the value of the character code must be less than 256 in non-UTF-8
193     mode, and less than 2**31 in UTF-8 mode (that is, the maximum hexadecimal value
194     is 7FFFFFFF). If characters other than hexadecimal digits appear between \x{
195     and }, or if there is no terminating }, this form of escape is not recognized.
196     Instead, the initial \x will be interpreted as a basic hexadecimal escape,
197     with no following digits, giving a character whose value is zero.
198 nigel 63 </P>
199     <P>
200     Characters whose value is less than 256 can be defined by either of the two
201 nigel 87 syntaxes for \x. There is no difference in the way they are handled. For
202     example, \xdc is exactly the same as \x{dc}.
203 nigel 63 </P>
204     <P>
205 nigel 91 After \0 up to two further octal digits are read. If there are fewer than two
206     digits, just those that are present are used. Thus the sequence \0\x\07
207     specifies two binary zeros followed by a BEL character (code value 7). Make
208     sure you supply two digits after the initial zero if the pattern character that
209     follows is itself an octal digit.
210 nigel 63 </P>
211     <P>
212     The handling of a backslash followed by a digit other than 0 is complicated.
213     Outside a character class, PCRE reads it and any following digits as a decimal
214     number. If the number is less than 10, or if there have been at least that many
215     previous capturing left parentheses in the expression, the entire sequence is
216     taken as a <i>back reference</i>. A description of how this works is given
217 nigel 75 <a href="#backreferences">later,</a>
218     following the discussion of
219     <a href="#subpattern">parenthesized subpatterns.</a>
220 nigel 63 </P>
221     <P>
222     Inside a character class, or if the decimal number is greater than 9 and there
223     have not been that many capturing subpatterns, PCRE re-reads up to three octal
224 nigel 93 digits following the backslash, and uses them to generate a data character. Any
225 nigel 91 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
226     character specified in octal must be less than \400. In UTF-8 mode, values up
227     to \777 are permitted. For example:
228 nigel 63 <pre>
229     \040 is another way of writing a space
230 nigel 75 \40 is the same, provided there are fewer than 40 previous capturing subpatterns
231 nigel 63 \7 is always a back reference
232 nigel 75 \11 might be a back reference, or another way of writing a tab
233 nigel 63 \011 is always a tab
234     \0113 is a tab followed by the character "3"
235 nigel 75 \113 might be a back reference, otherwise the character with octal code 113
236     \377 might be a back reference, otherwise the byte consisting entirely of 1 bits
237     \81 is either a back reference, or a binary zero followed by the two characters "8" and "1"
238     </pre>
239 nigel 63 Note that octal values of 100 or greater must not be introduced by a leading
240     zero, because no more than three octal digits are ever read.
241     </P>
242     <P>
243 nigel 91 All the sequences that define a single character value can be used both inside
244     and outside character classes. In addition, inside a character class, the
245     sequence \b is interpreted as the backspace character (hex 08), and the
246 nigel 93 sequences \R and \X are interpreted as the characters "R" and "X",
247     respectively. Outside a character class, these sequences have different
248     meanings
249 nigel 75 <a href="#uniextseq">(see below).</a>
250 nigel 63 </P>
251 nigel 75 <br><b>
252 nigel 93 Absolute and relative back references
253     </b><br>
254     <P>
255     The sequence \g followed by a positive or negative number, optionally enclosed
256     in braces, is an absolute or relative back reference. Back references are
257     discussed
258     <a href="#backreferences">later,</a>
259     following the discussion of
260     <a href="#subpattern">parenthesized subpatterns.</a>
261     </P>
262     <br><b>
263 nigel 75 Generic character types
264     </b><br>
265 nigel 63 <P>
266 nigel 93 Another use of backslash is for specifying generic character types. The
267 nigel 75 following are always recognized:
268 nigel 63 <pre>
269     \d any decimal digit
270     \D any character that is not a decimal digit
271     \s any whitespace character
272     \S any character that is not a whitespace character
273     \w any "word" character
274     \W any "non-word" character
275 nigel 75 </pre>
276 nigel 63 Each pair of escape sequences partitions the complete set of characters into
277     two disjoint sets. Any given character matches one, and only one, of each pair.
278     </P>
279     <P>
280 nigel 75 These character type sequences can appear both inside and outside character
281     classes. They each match one character of the appropriate type. If the current
282     matching point is at the end of the subject string, all of them fail, since
283     there is no character to match.
284 nigel 63 </P>
285     <P>
286     For compatibility with Perl, \s does not match the VT character (code 11).
287     This makes it different from the the POSIX "space" class. The \s characters
288 nigel 91 are HT (9), LF (10), FF (12), CR (13), and space (32). (If "use locale;" is
289     included in a Perl script, \s may match the VT character. In PCRE, it never
290     does.)
291 nigel 63 </P>
292     <P>
293 nigel 75 A "word" character is an underscore or any character less than 256 that is a
294     letter or digit. The definition of letters and digits is controlled by PCRE's
295     low-valued character tables, and may vary if locale-specific matching is taking
296     place (see
297 nigel 63 <a href="pcreapi.html#localesupport">"Locale support"</a>
298     in the
299     <a href="pcreapi.html"><b>pcreapi</b></a>
300 ph10 142 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
301     or "french" in Windows, some character codes greater than 128 are used for
302     accented letters, and these are matched by \w.
303 nigel 63 </P>
304     <P>
305 nigel 75 In UTF-8 mode, characters with values greater than 128 never match \d, \s, or
306     \w, and always match \D, \S, and \W. This is true even when Unicode
307 nigel 87 character property support is available. The use of locales with Unicode is
308     discouraged.
309 nigel 93 </P>
310     <br><b>
311     Newline sequences
312     </b><br>
313     <P>
314     Outside a character class, the escape sequence \R matches any Unicode newline
315     sequence. This is an extension to Perl. In non-UTF-8 mode \R is equivalent to
316     the following:
317     <pre>
318     (?&#62;\r\n|\n|\x0b|\f|\r|\x85)
319     </pre>
320     This is an example of an "atomic group", details of which are given
321     <a href="#atomicgroup">below.</a>
322     This particular group matches either the two-character sequence CR followed by
323     LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
324     U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
325     line, U+0085). The two-character sequence is treated as a single unit that
326     cannot be split.
327     </P>
328     <P>
329     In UTF-8 mode, two additional characters whose codepoints are greater than 255
330     are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
331     Unicode character property support is not needed for these characters to be
332     recognized.
333     </P>
334     <P>
335     Inside a character class, \R matches the letter "R".
336 nigel 75 <a name="uniextseq"></a></P>
337     <br><b>
338     Unicode character properties
339     </b><br>
340     <P>
341     When PCRE is built with Unicode character property support, three additional
342 nigel 87 escape sequences to match character properties are available when UTF-8 mode
343 nigel 75 is selected. They are:
344     <pre>
345 nigel 87 \p{<i>xx</i>} a character with the <i>xx</i> property
346     \P{<i>xx</i>} a character without the <i>xx</i> property
347     \X an extended Unicode sequence
348 nigel 75 </pre>
349 nigel 87 The property names represented by <i>xx</i> above are limited to the Unicode
350     script names, the general category properties, and "Any", which matches any
351     character (including newline). Other properties such as "InMusicalSymbols" are
352     not currently supported by PCRE. Note that \P{Any} does not match any
353     characters, so always causes a match failure.
354 nigel 63 </P>
355     <P>
356 nigel 87 Sets of Unicode characters are defined as belonging to certain scripts. A
357     character from one of these sets can be matched using a script name. For
358     example:
359 nigel 75 <pre>
360 nigel 87 \p{Greek}
361     \P{Han}
362     </pre>
363     Those that are not part of an identified script are lumped together as
364     "Common". The current list of scripts is:
365     </P>
366     <P>
367     Arabic,
368     Armenian,
369 nigel 93 Balinese,
370 nigel 87 Bengali,
371     Bopomofo,
372     Braille,
373     Buginese,
374     Buhid,
375     Canadian_Aboriginal,
376     Cherokee,
377     Common,
378     Coptic,
379 nigel 93 Cuneiform,
380 nigel 87 Cypriot,
381     Cyrillic,
382     Deseret,
383     Devanagari,
384     Ethiopic,
385     Georgian,
386     Glagolitic,
387     Gothic,
388     Greek,
389     Gujarati,
390     Gurmukhi,
391     Han,
392     Hangul,
393     Hanunoo,
394     Hebrew,
395     Hiragana,
396     Inherited,
397     Kannada,
398     Katakana,
399     Kharoshthi,
400     Khmer,
401     Lao,
402     Latin,
403     Limbu,
404     Linear_B,
405     Malayalam,
406     Mongolian,
407     Myanmar,
408     New_Tai_Lue,
409 nigel 93 Nko,
410 nigel 87 Ogham,
411     Old_Italic,
412     Old_Persian,
413     Oriya,
414     Osmanya,
415 nigel 93 Phags_Pa,
416     Phoenician,
417 nigel 87 Runic,
418     Shavian,
419     Sinhala,
420     Syloti_Nagri,
421     Syriac,
422     Tagalog,
423     Tagbanwa,
424     Tai_Le,
425     Tamil,
426     Telugu,
427     Thaana,
428     Thai,
429     Tibetan,
430     Tifinagh,
431     Ugaritic,
432     Yi.
433     </P>
434     <P>
435     Each character has exactly one general category property, specified by a
436     two-letter abbreviation. For compatibility with Perl, negation can be specified
437     by including a circumflex between the opening brace and the property name. For
438     example, \p{^Lu} is the same as \P{Lu}.
439     </P>
440     <P>
441     If only one letter is specified with \p or \P, it includes all the general
442     category properties that start with that letter. In this case, in the absence
443     of negation, the curly brackets in the escape sequence are optional; these two
444     examples have the same effect:
445     <pre>
446 nigel 75 \p{L}
447     \pL
448     </pre>
449 nigel 87 The following general category property codes are supported:
450 nigel 75 <pre>
451     C Other
452     Cc Control
453     Cf Format
454     Cn Unassigned
455     Co Private use
456     Cs Surrogate
457    
458     L Letter
459     Ll Lower case letter
460     Lm Modifier letter
461     Lo Other letter
462     Lt Title case letter
463     Lu Upper case letter
464    
465     M Mark
466     Mc Spacing mark
467     Me Enclosing mark
468     Mn Non-spacing mark
469    
470     N Number
471     Nd Decimal number
472     Nl Letter number
473     No Other number
474    
475     P Punctuation
476     Pc Connector punctuation
477     Pd Dash punctuation
478     Pe Close punctuation
479     Pf Final punctuation
480     Pi Initial punctuation
481     Po Other punctuation
482     Ps Open punctuation
483    
484     S Symbol
485     Sc Currency symbol
486     Sk Modifier symbol
487     Sm Mathematical symbol
488     So Other symbol
489    
490     Z Separator
491     Zl Line separator
492     Zp Paragraph separator
493     Zs Space separator
494     </pre>
495 nigel 87 The special property L& is also supported: it matches a character that has
496     the Lu, Ll, or Lt property, in other words, a letter that is not classified as
497     a modifier or "other".
498 nigel 75 </P>
499     <P>
500 nigel 87 The long synonyms for these properties that Perl supports (such as \p{Letter})
501 nigel 91 are not supported by PCRE, nor is it permitted to prefix any of these
502 nigel 87 properties with "Is".
503     </P>
504     <P>
505     No character that is in the Unicode table has the Cn (unassigned) property.
506     Instead, this property is assumed for any code point that is not in the
507     Unicode table.
508     </P>
509     <P>
510 nigel 75 Specifying caseless matching does not affect these escape sequences. For
511     example, \p{Lu} always matches only upper case letters.
512     </P>
513     <P>
514     The \X escape matches any number of Unicode characters that form an extended
515     Unicode sequence. \X is equivalent to
516     <pre>
517     (?&#62;\PM\pM*)
518     </pre>
519     That is, it matches a character without the "mark" property, followed by zero
520     or more characters with the "mark" property, and treats the sequence as an
521     atomic group
522     <a href="#atomicgroup">(see below).</a>
523     Characters with the "mark" property are typically accents that affect the
524     preceding character.
525     </P>
526     <P>
527     Matching characters by Unicode property is not fast, because PCRE has to search
528     a structure that contains data for over fifteen thousand characters. That is
529     why the traditional escape sequences such as \d and \w do not use Unicode
530     properties in PCRE.
531     <a name="smallassertions"></a></P>
532     <br><b>
533     Simple assertions
534     </b><br>
535     <P>
536 nigel 93 The final use of backslash is for certain simple assertions. An assertion
537 nigel 63 specifies a condition that has to be met at a particular point in a match,
538     without consuming any characters from the subject string. The use of
539 nigel 75 subpatterns for more complicated assertions is described
540     <a href="#bigassertions">below.</a>
541 nigel 91 The backslashed assertions are:
542 nigel 63 <pre>
543     \b matches at a word boundary
544     \B matches when not at a word boundary
545 nigel 93 \A matches at the start of the subject
546     \Z matches at the end of the subject
547     also matches before a newline at the end of the subject
548     \z matches only at the end of the subject
549     \G matches at the first matching position in the subject
550 nigel 75 </pre>
551 nigel 63 These assertions may not appear in character classes (but note that \b has a
552     different meaning, namely the backspace character, inside a character class).
553     </P>
554     <P>
555     A word boundary is a position in the subject string where the current character
556     and the previous character do not both match \w or \W (i.e. one matches
557     \w and the other matches \W), or the start or end of the string if the
558     first or last character matches \w, respectively.
559     </P>
560     <P>
561     The \A, \Z, and \z assertions differ from the traditional circumflex and
562 nigel 75 dollar (described in the next section) in that they only ever match at the very
563     start and end of the subject string, whatever options are set. Thus, they are
564     independent of multiline mode. These three assertions are not affected by the
565     PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
566     circumflex and dollar metacharacters. However, if the <i>startoffset</i>
567     argument of <b>pcre_exec()</b> is non-zero, indicating that matching is to start
568     at a point other than the beginning of the subject, \A can never match. The
569 nigel 91 difference between \Z and \z is that \Z matches before a newline at the end
570     of the string as well as at the very end, whereas \z matches only at the end.
571 nigel 63 </P>
572     <P>
573     The \G assertion is true only when the current matching position is at the
574     start point of the match, as specified by the <i>startoffset</i> argument of
575     <b>pcre_exec()</b>. It differs from \A when the value of <i>startoffset</i> is
576     non-zero. By calling <b>pcre_exec()</b> multiple times with appropriate
577     arguments, you can mimic Perl's /g option, and it is in this kind of
578     implementation where \G can be useful.
579     </P>
580     <P>
581     Note, however, that PCRE's interpretation of \G, as the start of the current
582     match, is subtly different from Perl's, which defines it as the end of the
583     previous match. In Perl, these can be different when the previously matched
584     string was empty. Because PCRE does just one match at a time, it cannot
585     reproduce this behaviour.
586     </P>
587     <P>
588     If all the alternatives of a pattern begin with \G, the expression is anchored
589     to the starting match position, and the "anchored" flag is set in the compiled
590     regular expression.
591     </P>
592 nigel 93 <br><a name="SEC4" href="#TOC1">CIRCUMFLEX AND DOLLAR</a><br>
593 nigel 63 <P>
594     Outside a character class, in the default matching mode, the circumflex
595 nigel 75 character is an assertion that is true only if the current matching point is
596 nigel 63 at the start of the subject string. If the <i>startoffset</i> argument of
597     <b>pcre_exec()</b> is non-zero, circumflex can never match if the PCRE_MULTILINE
598     option is unset. Inside a character class, circumflex has an entirely different
599 nigel 75 meaning
600     <a href="#characterclass">(see below).</a>
601 nigel 63 </P>
602     <P>
603     Circumflex need not be the first character of the pattern if a number of
604     alternatives are involved, but it should be the first thing in each alternative
605     in which it appears if the pattern is ever to match that branch. If all
606     possible alternatives start with a circumflex, that is, if the pattern is
607     constrained to match only at the start of the subject, it is said to be an
608     "anchored" pattern. (There are also other constructs that can cause a pattern
609     to be anchored.)
610     </P>
611     <P>
612 nigel 75 A dollar character is an assertion that is true only if the current matching
613 nigel 63 point is at the end of the subject string, or immediately before a newline
614 nigel 91 at the end of the string (by default). Dollar need not be the last character of
615     the pattern if a number of alternatives are involved, but it should be the last
616     item in any branch in which it appears. Dollar has no special meaning in a
617     character class.
618 nigel 63 </P>
619     <P>
620     The meaning of dollar can be changed so that it matches only at the very end of
621     the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
622     does not affect the \Z assertion.
623     </P>
624     <P>
625     The meanings of the circumflex and dollar characters are changed if the
626 nigel 91 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
627     immediately after internal newlines as well as at the start of the subject
628     string. It does not match after a newline that ends the string. A dollar
629     matches before any newlines in the string, as well as at the very end, when
630     PCRE_MULTILINE is set. When newline is specified as the two-character
631     sequence CRLF, isolated CR and LF characters do not indicate newlines.
632 nigel 63 </P>
633     <P>
634 nigel 91 For example, the pattern /^abc$/ matches the subject string "def\nabc" (where
635     \n represents a newline) in multiline mode, but not otherwise. Consequently,
636     patterns that are anchored in single line mode because all branches start with
637     ^ are not anchored in multiline mode, and a match for circumflex is possible
638     when the <i>startoffset</i> argument of <b>pcre_exec()</b> is non-zero. The
639     PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
640     </P>
641     <P>
642 nigel 63 Note that the sequences \A, \Z, and \z can be used to match the start and
643     end of the subject in both modes, and if all branches of a pattern start with
644 nigel 91 \A it is always anchored, whether or not PCRE_MULTILINE is set.
645 nigel 63 </P>
646 nigel 93 <br><a name="SEC5" href="#TOC1">FULL STOP (PERIOD, DOT)</a><br>
647 nigel 63 <P>
648     Outside a character class, a dot in the pattern matches any one character in
649 nigel 91 the subject string except (by default) a character that signifies the end of a
650 nigel 93 line. In UTF-8 mode, the matched character may be more than one byte long.
651 nigel 63 </P>
652 nigel 91 <P>
653 nigel 93 When a line ending is defined as a single character, dot never matches that
654     character; when the two-character sequence CRLF is used, dot does not match CR
655     if it is immediately followed by LF, but otherwise it matches all characters
656     (including isolated CRs and LFs). When any Unicode line endings are being
657     recognized, dot does not match CR or LF or any of the other line ending
658     characters.
659     </P>
660     <P>
661 nigel 91 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
662 nigel 93 option is set, a dot matches any one character, without exception. If the
663     two-character sequence CRLF is present in the subject string, it takes two dots
664     to match it.
665 nigel 91 </P>
666     <P>
667     The handling of dot is entirely independent of the handling of circumflex and
668     dollar, the only relationship being that they both involve newlines. Dot has no
669     special meaning in a character class.
670     </P>
671 nigel 93 <br><a name="SEC6" href="#TOC1">MATCHING A SINGLE BYTE</a><br>
672 nigel 63 <P>
673     Outside a character class, the escape sequence \C matches any one byte, both
674 nigel 93 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
675     characters. The feature is provided in Perl in order to match individual bytes
676     in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes,
677     what remains in the string may be a malformed UTF-8 string. For this reason,
678     the \C escape sequence is best avoided.
679 nigel 63 </P>
680     <P>
681 nigel 75 PCRE does not allow \C to appear in lookbehind assertions
682     <a href="#lookbehind">(described below),</a>
683     because in UTF-8 mode this would make it impossible to calculate the length of
684     the lookbehind.
685     <a name="characterclass"></a></P>
686 nigel 93 <br><a name="SEC7" href="#TOC1">SQUARE BRACKETS AND CHARACTER CLASSES</a><br>
687 nigel 63 <P>
688     An opening square bracket introduces a character class, terminated by a closing
689     square bracket. A closing square bracket on its own is not special. If a
690     closing square bracket is required as a member of the class, it should be the
691     first data character in the class (after an initial circumflex, if present) or
692     escaped with a backslash.
693     </P>
694     <P>
695     A character class matches a single character in the subject. In UTF-8 mode, the
696     character may occupy more than one byte. A matched character must be in the set
697     of characters defined by the class, unless the first character in the class
698     definition is a circumflex, in which case the subject character must not be in
699     the set defined by the class. If a circumflex is actually required as a member
700     of the class, ensure it is not the first character, or escape it with a
701     backslash.
702     </P>
703     <P>
704     For example, the character class [aeiou] matches any lower case vowel, while
705     [^aeiou] matches any character that is not a lower case vowel. Note that a
706 nigel 75 circumflex is just a convenient notation for specifying the characters that
707     are in the class by enumerating those that are not. A class that starts with a
708     circumflex is not an assertion: it still consumes a character from the subject
709     string, and therefore it fails if the current pointer is at the end of the
710     string.
711 nigel 63 </P>
712     <P>
713     In UTF-8 mode, characters with values greater than 255 can be included in a
714     class as a literal string of bytes, or by using the \x{ escaping mechanism.
715     </P>
716     <P>
717     When caseless matching is set, any letters in a class represent both their
718     upper case and lower case versions, so for example, a caseless [aeiou] matches
719     "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
720 nigel 77 caseful version would. In UTF-8 mode, PCRE always understands the concept of
721     case for characters whose values are less than 128, so caseless matching is
722     always possible. For characters with higher values, the concept of case is
723     supported if PCRE is compiled with Unicode property support, but not otherwise.
724     If you want to use caseless matching for characters 128 and above, you must
725     ensure that PCRE is compiled with Unicode property support as well as with
726     UTF-8 support.
727 nigel 63 </P>
728     <P>
729 nigel 93 Characters that might indicate line breaks are never treated in any special way
730     when matching character classes, whatever line-ending sequence is in use, and
731     whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
732     such as [^a] always matches one of these characters.
733 nigel 63 </P>
734     <P>
735     The minus (hyphen) character can be used to specify a range of characters in a
736     character class. For example, [d-m] matches any letter between d and m,
737     inclusive. If a minus character is required in a class, it must be escaped with
738     a backslash or appear in a position where it cannot be interpreted as
739     indicating a range, typically as the first or last character in the class.
740     </P>
741     <P>
742     It is not possible to have the literal character "]" as the end character of a
743     range. A pattern such as [W-]46] is interpreted as a class of two characters
744     ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
745     "-46]". However, if the "]" is escaped with a backslash it is interpreted as
746 nigel 75 the end of range, so [W-\]46] is interpreted as a class containing a range
747     followed by two other characters. The octal or hexadecimal representation of
748     "]" can also be used to end a range.
749 nigel 63 </P>
750     <P>
751     Ranges operate in the collating sequence of character values. They can also be
752     used for characters specified numerically, for example [\000-\037]. In UTF-8
753     mode, ranges can include characters whose values are greater than 255, for
754     example [\x{100}-\x{2ff}].
755     </P>
756     <P>
757     If a range that includes letters is used when caseless matching is set, it
758     matches the letters in either case. For example, [W-c] is equivalent to
759 nigel 75 [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
760 ph10 142 tables for a French locale are in use, [\xc8-\xcb] matches accented E
761 nigel 75 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
762     characters with values greater than 128 only when it is compiled with Unicode
763     property support.
764 nigel 63 </P>
765     <P>
766 nigel 75 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
767     in a character class, and add the characters that they match to the class. For
768 nigel 63 example, [\dABCDEF] matches any hexadecimal digit. A circumflex can
769     conveniently be used with the upper case character types to specify a more
770     restricted set of characters than the matching lower case type. For example,
771     the class [^\W_] matches any letter or digit, but not underscore.
772     </P>
773     <P>
774 nigel 75 The only metacharacters that are recognized in character classes are backslash,
775     hyphen (only where it can be interpreted as specifying a range), circumflex
776     (only at the start), opening square bracket (only when it can be interpreted as
777     introducing a POSIX class name - see the next section), and the terminating
778     closing square bracket. However, escaping other non-alphanumeric characters
779     does no harm.
780 nigel 63 </P>
781 nigel 93 <br><a name="SEC8" href="#TOC1">POSIX CHARACTER CLASSES</a><br>
782 nigel 63 <P>
783 nigel 75 Perl supports the POSIX notation for character classes. This uses names
784 nigel 63 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
785     this notation. For example,
786     <pre>
787     [01[:alpha:]%]
788 nigel 75 </pre>
789 nigel 63 matches "0", "1", any alphabetic character, or "%". The supported class names
790     are
791     <pre>
792     alnum letters and digits
793     alpha letters
794     ascii character codes 0 - 127
795     blank space or tab only
796     cntrl control characters
797     digit decimal digits (same as \d)
798     graph printing characters, excluding space
799     lower lower case letters
800     print printing characters, including space
801     punct printing characters, excluding letters and digits
802     space white space (not quite the same as \s)
803     upper upper case letters
804     word "word" characters (same as \w)
805     xdigit hexadecimal digits
806 nigel 75 </pre>
807 nigel 63 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
808     space (32). Notice that this list includes the VT character (code 11). This
809     makes "space" different to \s, which does not include VT (for Perl
810     compatibility).
811     </P>
812     <P>
813     The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
814     5.8. Another Perl extension is negation, which is indicated by a ^ character
815     after the colon. For example,
816     <pre>
817     [12[:^digit:]]
818 nigel 75 </pre>
819 nigel 63 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
820     syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
821     supported, and an error is given if they are encountered.
822     </P>
823     <P>
824 nigel 75 In UTF-8 mode, characters with values greater than 128 do not match any of
825 nigel 63 the POSIX character classes.
826     </P>
827 nigel 93 <br><a name="SEC9" href="#TOC1">VERTICAL BAR</a><br>
828 nigel 63 <P>
829     Vertical bar characters are used to separate alternative patterns. For example,
830     the pattern
831     <pre>
832     gilbert|sullivan
833 nigel 75 </pre>
834 nigel 63 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
835 nigel 91 and an empty alternative is permitted (matching the empty string). The matching
836     process tries each alternative in turn, from left to right, and the first one
837     that succeeds is used. If the alternatives are within a subpattern
838 nigel 75 <a href="#subpattern">(defined below),</a>
839     "succeeds" means matching the rest of the main pattern as well as the
840     alternative in the subpattern.
841 nigel 63 </P>
842 nigel 93 <br><a name="SEC10" href="#TOC1">INTERNAL OPTION SETTING</a><br>
843 nigel 63 <P>
844     The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
845     PCRE_EXTENDED options can be changed from within the pattern by a sequence of
846     Perl option letters enclosed between "(?" and ")". The option letters are
847     <pre>
848     i for PCRE_CASELESS
849     m for PCRE_MULTILINE
850     s for PCRE_DOTALL
851     x for PCRE_EXTENDED
852 nigel 75 </pre>
853 nigel 63 For example, (?im) sets caseless, multiline matching. It is also possible to
854     unset these options by preceding the letter with a hyphen, and a combined
855     setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
856     PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
857     permitted. If a letter appears both before and after the hyphen, the option is
858     unset.
859     </P>
860     <P>
861     When an option change occurs at top level (that is, not inside subpattern
862     parentheses), the change applies to the remainder of the pattern that follows.
863     If the change is placed right at the start of a pattern, PCRE extracts it into
864     the global options (and it will therefore show up in data extracted by the
865     <b>pcre_fullinfo()</b> function).
866     </P>
867     <P>
868 nigel 93 An option change within a subpattern (see below for a description of
869     subpatterns) affects only that part of the current pattern that follows it, so
870 nigel 63 <pre>
871     (a(?i)b)c
872 nigel 75 </pre>
873 nigel 63 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
874     By this means, options can be made to have different settings in different
875     parts of the pattern. Any changes made in one alternative do carry on
876     into subsequent branches within the same subpattern. For example,
877     <pre>
878     (a(?i)b|c)
879 nigel 75 </pre>
880 nigel 63 matches "ab", "aB", "c", and "C", even though when matching "C" the first
881     branch is abandoned before the option setting. This is because the effects of
882     option settings happen at compile time. There would be some very weird
883     behaviour otherwise.
884     </P>
885     <P>
886 nigel 91 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
887     changed in the same way as the Perl-compatible options by using the characters
888     J, U and X respectively.
889 nigel 75 <a name="subpattern"></a></P>
890 nigel 93 <br><a name="SEC11" href="#TOC1">SUBPATTERNS</a><br>
891 nigel 63 <P>
892     Subpatterns are delimited by parentheses (round brackets), which can be nested.
893 nigel 75 Turning part of a pattern into a subpattern does two things:
894     <br>
895     <br>
896 nigel 63 1. It localizes a set of alternatives. For example, the pattern
897     <pre>
898     cat(aract|erpillar|)
899 nigel 75 </pre>
900 nigel 63 matches one of the words "cat", "cataract", or "caterpillar". Without the
901 nigel 93 parentheses, it would match "cataract", "erpillar" or an empty string.
902 nigel 75 <br>
903     <br>
904     2. It sets up the subpattern as a capturing subpattern. This means that, when
905     the whole pattern matches, that portion of the subject string that matched the
906     subpattern is passed back to the caller via the <i>ovector</i> argument of
907 nigel 63 <b>pcre_exec()</b>. Opening parentheses are counted from left to right (starting
908 nigel 75 from 1) to obtain numbers for the capturing subpatterns.
909 nigel 63 </P>
910     <P>
911     For example, if the string "the red king" is matched against the pattern
912     <pre>
913     the ((red|white) (king|queen))
914 nigel 75 </pre>
915 nigel 63 the captured substrings are "red king", "red", and "king", and are numbered 1,
916     2, and 3, respectively.
917     </P>
918     <P>
919     The fact that plain parentheses fulfil two functions is not always helpful.
920     There are often times when a grouping subpattern is required without a
921     capturing requirement. If an opening parenthesis is followed by a question mark
922     and a colon, the subpattern does not do any capturing, and is not counted when
923     computing the number of any subsequent capturing subpatterns. For example, if
924     the string "the white queen" is matched against the pattern
925     <pre>
926     the ((?:red|white) (king|queen))
927 nigel 75 </pre>
928 nigel 63 the captured substrings are "white queen" and "queen", and are numbered 1 and
929 nigel 93 2. The maximum number of capturing subpatterns is 65535.
930 nigel 63 </P>
931     <P>
932     As a convenient shorthand, if any option settings are required at the start of
933     a non-capturing subpattern, the option letters may appear between the "?" and
934     the ":". Thus the two patterns
935     <pre>
936     (?i:saturday|sunday)
937     (?:(?i)saturday|sunday)
938 nigel 75 </pre>
939 nigel 63 match exactly the same set of strings. Because alternative branches are tried
940     from left to right, and options are not reset until the end of the subpattern
941     is reached, an option setting in one branch does affect subsequent branches, so
942     the above patterns match "SUNDAY" as well as "Saturday".
943     </P>
944 nigel 93 <br><a name="SEC12" href="#TOC1">NAMED SUBPATTERNS</a><br>
945 nigel 63 <P>
946     Identifying capturing parentheses by number is simple, but it can be very hard
947     to keep track of the numbers in complicated regular expressions. Furthermore,
948 nigel 75 if an expression is modified, the numbers may change. To help with this
949 nigel 93 difficulty, PCRE supports the naming of subpatterns. This feature was not
950     added to Perl until release 5.10. Python had the feature earlier, and PCRE
951     introduced it at release 4.0, using the Python syntax. PCRE now supports both
952     the Perl and the Python syntax.
953     </P>
954     <P>
955     In PCRE, a subpattern can be named in one of three ways: (?&#60;name&#62;...) or
956     (?'name'...) as in Perl, or (?P&#60;name&#62;...) as in Python. References to capturing
957 nigel 91 parentheses from other parts of the pattern, such as
958     <a href="#backreferences">backreferences,</a>
959     <a href="#recursion">recursion,</a>
960     and
961     <a href="#conditions">conditions,</a>
962     can be made by name as well as by number.
963 nigel 63 </P>
964     <P>
965 nigel 91 Names consist of up to 32 alphanumeric characters and underscores. Named
966 nigel 93 capturing parentheses are still allocated numbers as well as names, exactly as
967     if the names were not present. The PCRE API provides function calls for
968     extracting the name-to-number translation table from a compiled pattern. There
969     is also a convenience function for extracting a captured substring by name.
970 nigel 91 </P>
971     <P>
972     By default, a name must be unique within a pattern, but it is possible to relax
973     this constraint by setting the PCRE_DUPNAMES option at compile time. This can
974     be useful for patterns where only one instance of the named parentheses can
975     match. Suppose you want to match the name of a weekday, either as a 3-letter
976     abbreviation or as the full name, and in both cases you want to extract the
977     abbreviation. This pattern (ignoring the line breaks) does the job:
978     <pre>
979 nigel 93 (?&#60;DN&#62;Mon|Fri|Sun)(?:day)?|
980     (?&#60;DN&#62;Tue)(?:sday)?|
981     (?&#60;DN&#62;Wed)(?:nesday)?|
982     (?&#60;DN&#62;Thu)(?:rsday)?|
983     (?&#60;DN&#62;Sat)(?:urday)?
984 nigel 91 </pre>
985     There are five capturing substrings, but only one is ever set after a match.
986     The convenience function for extracting the data by name returns the substring
987 nigel 93 for the first (and in this example, the only) subpattern of that name that
988 nigel 91 matched. This saves searching to find which numbered subpattern it was. If you
989     make a reference to a non-unique named subpattern from elsewhere in the
990     pattern, the one that corresponds to the lowest number is used. For further
991     details of the interfaces for handling named subpatterns, see the
992 nigel 63 <a href="pcreapi.html"><b>pcreapi</b></a>
993     documentation.
994     </P>
995 nigel 93 <br><a name="SEC13" href="#TOC1">REPETITION</a><br>
996 nigel 63 <P>
997     Repetition is specified by quantifiers, which can follow any of the following
998     items:
999     <pre>
1000     a literal data character
1001 nigel 93 the dot metacharacter
1002 nigel 63 the \C escape sequence
1003 nigel 75 the \X escape sequence (in UTF-8 mode with Unicode properties)
1004 nigel 93 the \R escape sequence
1005 nigel 75 an escape such as \d that matches a single character
1006 nigel 63 a character class
1007     a back reference (see next section)
1008     a parenthesized subpattern (unless it is an assertion)
1009 nigel 75 </pre>
1010 nigel 63 The general repetition quantifier specifies a minimum and maximum number of
1011     permitted matches, by giving the two numbers in curly brackets (braces),
1012     separated by a comma. The numbers must be less than 65536, and the first must
1013     be less than or equal to the second. For example:
1014     <pre>
1015     z{2,4}
1016 nigel 75 </pre>
1017 nigel 63 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1018     character. If the second number is omitted, but the comma is present, there is
1019     no upper limit; if the second number and the comma are both omitted, the
1020     quantifier specifies an exact number of required matches. Thus
1021     <pre>
1022     [aeiou]{3,}
1023 nigel 75 </pre>
1024 nigel 63 matches at least 3 successive vowels, but may match many more, while
1025     <pre>
1026     \d{8}
1027 nigel 75 </pre>
1028 nigel 63 matches exactly 8 digits. An opening curly bracket that appears in a position
1029     where a quantifier is not allowed, or one that does not match the syntax of a
1030     quantifier, is taken as a literal character. For example, {,6} is not a
1031     quantifier, but a literal string of four characters.
1032     </P>
1033     <P>
1034     In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1035     bytes. Thus, for example, \x{100}{2} matches two UTF-8 characters, each of
1036 nigel 75 which is represented by a two-byte sequence. Similarly, when Unicode property
1037     support is available, \X{3} matches three Unicode extended sequences, each of
1038     which may be several bytes long (and they may be of different lengths).
1039 nigel 63 </P>
1040     <P>
1041     The quantifier {0} is permitted, causing the expression to behave as if the
1042     previous item and the quantifier were not present.
1043     </P>
1044     <P>
1045 nigel 93 For convenience, the three most common quantifiers have single-character
1046     abbreviations:
1047 nigel 63 <pre>
1048     * is equivalent to {0,}
1049     + is equivalent to {1,}
1050     ? is equivalent to {0,1}
1051 nigel 75 </pre>
1052 nigel 63 It is possible to construct infinite loops by following a subpattern that can
1053     match no characters with a quantifier that has no upper limit, for example:
1054     <pre>
1055     (a?)*
1056 nigel 75 </pre>
1057 nigel 63 Earlier versions of Perl and PCRE used to give an error at compile time for
1058     such patterns. However, because there are cases where this can be useful, such
1059     patterns are now accepted, but if any repetition of the subpattern does in fact
1060     match no characters, the loop is forcibly broken.
1061     </P>
1062     <P>
1063     By default, the quantifiers are "greedy", that is, they match as much as
1064     possible (up to the maximum number of permitted times), without causing the
1065     rest of the pattern to fail. The classic example of where this gives problems
1066 nigel 75 is in trying to match comments in C programs. These appear between /* and */
1067     and within the comment, individual * and / characters may appear. An attempt to
1068     match C comments by applying the pattern
1069 nigel 63 <pre>
1070     /\*.*\*/
1071 nigel 75 </pre>
1072 nigel 63 to the string
1073     <pre>
1074 nigel 75 /* first comment */ not comment /* second comment */
1075     </pre>
1076 nigel 63 fails, because it matches the entire string owing to the greediness of the .*
1077     item.
1078     </P>
1079     <P>
1080     However, if a quantifier is followed by a question mark, it ceases to be
1081     greedy, and instead matches the minimum number of times possible, so the
1082     pattern
1083     <pre>
1084     /\*.*?\*/
1085 nigel 75 </pre>
1086 nigel 63 does the right thing with the C comments. The meaning of the various
1087     quantifiers is not otherwise changed, just the preferred number of matches.
1088     Do not confuse this use of question mark with its use as a quantifier in its
1089     own right. Because it has two uses, it can sometimes appear doubled, as in
1090     <pre>
1091     \d??\d
1092 nigel 75 </pre>
1093 nigel 63 which matches one digit by preference, but can match two if that is the only
1094     way the rest of the pattern matches.
1095     </P>
1096     <P>
1097 nigel 93 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1098 nigel 63 the quantifiers are not greedy by default, but individual ones can be made
1099     greedy by following them with a question mark. In other words, it inverts the
1100     default behaviour.
1101     </P>
1102     <P>
1103     When a parenthesized subpattern is quantified with a minimum repeat count that
1104 nigel 75 is greater than 1 or with a limited maximum, more memory is required for the
1105 nigel 63 compiled pattern, in proportion to the size of the minimum or maximum.
1106     </P>
1107     <P>
1108     If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1109 nigel 93 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1110 nigel 63 implicitly anchored, because whatever follows will be tried against every
1111     character position in the subject string, so there is no point in retrying the
1112     overall match at any position after the first. PCRE normally treats such a
1113     pattern as though it were preceded by \A.
1114     </P>
1115     <P>
1116     In cases where it is known that the subject string contains no newlines, it is
1117     worth setting PCRE_DOTALL in order to obtain this optimization, or
1118     alternatively using ^ to indicate anchoring explicitly.
1119     </P>
1120     <P>
1121     However, there is one situation where the optimization cannot be used. When .*
1122     is inside capturing parentheses that are the subject of a backreference
1123 nigel 93 elsewhere in the pattern, a match at the start may fail where a later one
1124     succeeds. Consider, for example:
1125 nigel 63 <pre>
1126     (.*)abc\1
1127 nigel 75 </pre>
1128 nigel 63 If the subject is "xyz123abc123" the match point is the fourth character. For
1129     this reason, such a pattern is not implicitly anchored.
1130     </P>
1131     <P>
1132     When a capturing subpattern is repeated, the value captured is the substring
1133     that matched the final iteration. For example, after
1134     <pre>
1135     (tweedle[dume]{3}\s*)+
1136 nigel 75 </pre>
1137 nigel 63 has matched "tweedledum tweedledee" the value of the captured substring is
1138     "tweedledee". However, if there are nested capturing subpatterns, the
1139     corresponding captured values may have been set in previous iterations. For
1140     example, after
1141     <pre>
1142     /(a|(b))+/
1143 nigel 75 </pre>
1144 nigel 63 matches "aba" the value of the second captured substring is "b".
1145 nigel 75 <a name="atomicgroup"></a></P>
1146 nigel 93 <br><a name="SEC14" href="#TOC1">ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS</a><br>
1147 nigel 63 <P>
1148 nigel 93 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1149     repetition, failure of what follows normally causes the repeated item to be
1150     re-evaluated to see if a different number of repeats allows the rest of the
1151     pattern to match. Sometimes it is useful to prevent this, either to change the
1152     nature of the match, or to cause it fail earlier than it otherwise might, when
1153     the author of the pattern knows there is no point in carrying on.
1154 nigel 63 </P>
1155     <P>
1156     Consider, for example, the pattern \d+foo when applied to the subject line
1157     <pre>
1158     123456bar
1159 nigel 75 </pre>
1160 nigel 63 After matching all 6 digits and then failing to match "foo", the normal
1161     action of the matcher is to try again with only 5 digits matching the \d+
1162     item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1163     (a term taken from Jeffrey Friedl's book) provides the means for specifying
1164     that once a subpattern has matched, it is not to be re-evaluated in this way.
1165     </P>
1166     <P>
1167 nigel 93 If we use atomic grouping for the previous example, the matcher gives up
1168 nigel 63 immediately on failing to match "foo" the first time. The notation is a kind of
1169     special parenthesis, starting with (?&#62; as in this example:
1170     <pre>
1171 nigel 73 (?&#62;\d+)foo
1172 nigel 75 </pre>
1173 nigel 63 This kind of parenthesis "locks up" the part of the pattern it contains once
1174     it has matched, and a failure further into the pattern is prevented from
1175     backtracking into it. Backtracking past it to previous items, however, works as
1176     normal.
1177     </P>
1178     <P>
1179     An alternative description is that a subpattern of this type matches the string
1180     of characters that an identical standalone pattern would match, if anchored at
1181     the current point in the subject string.
1182     </P>
1183     <P>
1184     Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1185     the above example can be thought of as a maximizing repeat that must swallow
1186     everything it can. So, while both \d+ and \d+? are prepared to adjust the
1187     number of digits they match in order to make the rest of the pattern match,
1188     (?&#62;\d+) can only match an entire sequence of digits.
1189     </P>
1190     <P>
1191     Atomic groups in general can of course contain arbitrarily complicated
1192     subpatterns, and can be nested. However, when the subpattern for an atomic
1193     group is just a single repeated item, as in the example above, a simpler
1194     notation, called a "possessive quantifier" can be used. This consists of an
1195     additional + character following a quantifier. Using this notation, the
1196     previous example can be rewritten as
1197     <pre>
1198 nigel 75 \d++foo
1199     </pre>
1200 nigel 63 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1201     option is ignored. They are a convenient notation for the simpler forms of
1202 nigel 93 atomic group. However, there is no difference in the meaning of a possessive
1203     quantifier and the equivalent atomic group, though there may be a performance
1204     difference; possessive quantifiers should be slightly faster.
1205 nigel 63 </P>
1206     <P>
1207 nigel 93 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1208     Jeffrey Friedl originated the idea (and the name) in the first edition of his
1209     book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1210     package, and PCRE copied it from there. It ultimately found its way into Perl
1211     at release 5.10.
1212 nigel 63 </P>
1213     <P>
1214 nigel 93 PCRE has an optimization that automatically "possessifies" certain simple
1215     pattern constructs. For example, the sequence A+B is treated as A++B because
1216     there is no point in backtracking into a sequence of A's when B must follow.
1217     </P>
1218     <P>
1219 nigel 63 When a pattern contains an unlimited repeat inside a subpattern that can itself
1220     be repeated an unlimited number of times, the use of an atomic group is the
1221     only way to avoid some failing matches taking a very long time indeed. The
1222     pattern
1223     <pre>
1224     (\D+|&#60;\d+&#62;)*[!?]
1225 nigel 75 </pre>
1226 nigel 63 matches an unlimited number of substrings that either consist of non-digits, or
1227     digits enclosed in &#60;&#62;, followed by either ! or ?. When it matches, it runs
1228     quickly. However, if it is applied to
1229     <pre>
1230     aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1231 nigel 75 </pre>
1232 nigel 63 it takes a long time before reporting failure. This is because the string can
1233 nigel 75 be divided between the internal \D+ repeat and the external * repeat in a
1234     large number of ways, and all have to be tried. (The example uses [!?] rather
1235     than a single character at the end, because both PCRE and Perl have an
1236     optimization that allows for fast failure when a single character is used. They
1237     remember the last single character that is required for a match, and fail early
1238     if it is not present in the string.) If the pattern is changed so that it uses
1239     an atomic group, like this:
1240 nigel 63 <pre>
1241     ((?&#62;\D+)|&#60;\d+&#62;)*[!?]
1242 nigel 75 </pre>
1243 nigel 63 sequences of non-digits cannot be broken, and failure happens quickly.
1244 nigel 75 <a name="backreferences"></a></P>
1245 nigel 93 <br><a name="SEC15" href="#TOC1">BACK REFERENCES</a><br>
1246 nigel 63 <P>
1247     Outside a character class, a backslash followed by a digit greater than 0 (and
1248     possibly further digits) is a back reference to a capturing subpattern earlier
1249     (that is, to its left) in the pattern, provided there have been that many
1250     previous capturing left parentheses.
1251     </P>
1252     <P>
1253     However, if the decimal number following the backslash is less than 10, it is
1254     always taken as a back reference, and causes an error only if there are not
1255     that many capturing left parentheses in the entire pattern. In other words, the
1256     parentheses that are referenced need not be to the left of the reference for
1257 nigel 91 numbers less than 10. A "forward back reference" of this type can make sense
1258     when a repetition is involved and the subpattern to the right has participated
1259     in an earlier iteration.
1260     </P>
1261     <P>
1262 nigel 93 It is not possible to have a numerical "forward back reference" to a subpattern
1263     whose number is 10 or more using this syntax because a sequence such as \50 is
1264     interpreted as a character defined in octal. See the subsection entitled
1265 nigel 91 "Non-printing characters"
1266 nigel 75 <a href="#digitsafterbackslash">above</a>
1267 nigel 93 for further details of the handling of digits following a backslash. There is
1268     no such problem when named parentheses are used. A back reference to any
1269     subpattern is possible using named parentheses (see below).
1270 nigel 63 </P>
1271     <P>
1272 nigel 93 Another way of avoiding the ambiguity inherent in the use of digits following a
1273     backslash is to use the \g escape sequence, which is a feature introduced in
1274     Perl 5.10. This escape must be followed by a positive or a negative number,
1275     optionally enclosed in braces. These examples are all identical:
1276     <pre>
1277     (ring), \1
1278     (ring), \g1
1279     (ring), \g{1}
1280     </pre>
1281     A positive number specifies an absolute reference without the ambiguity that is
1282     present in the older syntax. It is also useful when literal digits follow the
1283     reference. A negative number is a relative reference. Consider this example:
1284     <pre>
1285     (abc(def)ghi)\g{-1}
1286     </pre>
1287     The sequence \g{-1} is a reference to the most recently started capturing
1288     subpattern before \g, that is, is it equivalent to \2. Similarly, \g{-2}
1289     would be equivalent to \1. The use of relative references can be helpful in
1290     long patterns, and also in patterns that are created by joining together
1291     fragments that contain references within themselves.
1292     </P>
1293     <P>
1294 nigel 63 A back reference matches whatever actually matched the capturing subpattern in
1295     the current subject string, rather than anything matching the subpattern
1296     itself (see
1297     <a href="#subpatternsassubroutines">"Subpatterns as subroutines"</a>
1298     below for a way of doing that). So the pattern
1299     <pre>
1300     (sens|respons)e and \1ibility
1301 nigel 75 </pre>
1302 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1303     "sense and responsibility". If caseful matching is in force at the time of the
1304     back reference, the case of letters is relevant. For example,
1305     <pre>
1306     ((?i)rah)\s+\1
1307 nigel 75 </pre>
1308 nigel 63 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1309     capturing subpattern is matched caselessly.
1310     </P>
1311     <P>
1312 nigel 93 Back references to named subpatterns use the Perl syntax \k&#60;name&#62; or \k'name'
1313     or the Python syntax (?P=name). We could rewrite the above example in either of
1314     the following ways:
1315 nigel 63 <pre>
1316 nigel 93 (?&#60;p1&#62;(?i)rah)\s+\k&#60;p1&#62;
1317 nigel 91 (?P&#60;p1&#62;(?i)rah)\s+(?P=p1)
1318 nigel 75 </pre>
1319 nigel 91 A subpattern that is referenced by name may appear in the pattern before or
1320     after the reference.
1321     </P>
1322     <P>
1323 nigel 63 There may be more than one back reference to the same subpattern. If a
1324     subpattern has not actually been used in a particular match, any back
1325     references to it always fail. For example, the pattern
1326     <pre>
1327     (a|(bc))\2
1328 nigel 75 </pre>
1329 nigel 63 always fails if it starts to match "a" rather than "bc". Because there may be
1330     many capturing parentheses in a pattern, all digits following the backslash are
1331     taken as part of a potential back reference number. If the pattern continues
1332     with a digit character, some delimiter must be used to terminate the back
1333     reference. If the PCRE_EXTENDED option is set, this can be whitespace.
1334 nigel 75 Otherwise an empty comment (see
1335     <a href="#comments">"Comments"</a>
1336     below) can be used.
1337 nigel 63 </P>
1338     <P>
1339     A back reference that occurs inside the parentheses to which it refers fails
1340     when the subpattern is first used, so, for example, (a\1) never matches.
1341     However, such references can be useful inside repeated subpatterns. For
1342     example, the pattern
1343     <pre>
1344     (a|b\1)+
1345 nigel 75 </pre>
1346 nigel 63 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1347     the subpattern, the back reference matches the character string corresponding
1348     to the previous iteration. In order for this to work, the pattern must be such
1349     that the first iteration does not need to match the back reference. This can be
1350     done using alternation, as in the example above, or by a quantifier with a
1351     minimum of zero.
1352 nigel 75 <a name="bigassertions"></a></P>
1353 nigel 93 <br><a name="SEC16" href="#TOC1">ASSERTIONS</a><br>
1354 nigel 63 <P>
1355     An assertion is a test on the characters following or preceding the current
1356     matching point that does not actually consume any characters. The simple
1357 nigel 75 assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are described
1358     <a href="#smallassertions">above.</a>
1359     </P>
1360     <P>
1361 nigel 63 More complicated assertions are coded as subpatterns. There are two kinds:
1362     those that look ahead of the current position in the subject string, and those
1363 nigel 75 that look behind it. An assertion subpattern is matched in the normal way,
1364     except that it does not cause the current matching position to be changed.
1365 nigel 63 </P>
1366     <P>
1367 nigel 75 Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1368     because it makes no sense to assert the same thing several times. If any kind
1369     of assertion contains capturing subpatterns within it, these are counted for
1370     the purposes of numbering the capturing subpatterns in the whole pattern.
1371     However, substring capturing is carried out only for positive assertions,
1372     because it does not make sense for negative assertions.
1373 nigel 63 </P>
1374 nigel 75 <br><b>
1375     Lookahead assertions
1376     </b><br>
1377 nigel 63 <P>
1378 nigel 91 Lookahead assertions start with (?= for positive assertions and (?! for
1379     negative assertions. For example,
1380 nigel 63 <pre>
1381     \w+(?=;)
1382 nigel 75 </pre>
1383 nigel 63 matches a word followed by a semicolon, but does not include the semicolon in
1384     the match, and
1385     <pre>
1386     foo(?!bar)
1387 nigel 75 </pre>
1388 nigel 63 matches any occurrence of "foo" that is not followed by "bar". Note that the
1389     apparently similar pattern
1390     <pre>
1391     (?!foo)bar
1392 nigel 75 </pre>
1393 nigel 63 does not find an occurrence of "bar" that is preceded by something other than
1394     "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1395     (?!foo) is always true when the next three characters are "bar". A
1396 nigel 75 lookbehind assertion is needed to achieve the other effect.
1397 nigel 63 </P>
1398     <P>
1399     If you want to force a matching failure at some point in a pattern, the most
1400     convenient way to do it is with (?!) because an empty string always matches, so
1401     an assertion that requires there not to be an empty string must always fail.
1402 nigel 75 <a name="lookbehind"></a></P>
1403     <br><b>
1404     Lookbehind assertions
1405     </b><br>
1406 nigel 63 <P>
1407     Lookbehind assertions start with (?&#60;= for positive assertions and (?&#60;! for
1408     negative assertions. For example,
1409     <pre>
1410     (?&#60;!foo)bar
1411 nigel 75 </pre>
1412 nigel 63 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1413     a lookbehind assertion are restricted such that all the strings it matches must
1414 nigel 91 have a fixed length. However, if there are several top-level alternatives, they
1415     do not all have to have the same fixed length. Thus
1416 nigel 63 <pre>
1417     (?&#60;=bullock|donkey)
1418 nigel 75 </pre>
1419 nigel 63 is permitted, but
1420     <pre>
1421     (?&#60;!dogs?|cats?)
1422 nigel 75 </pre>
1423 nigel 63 causes an error at compile time. Branches that match different length strings
1424     are permitted only at the top level of a lookbehind assertion. This is an
1425     extension compared with Perl (at least for 5.8), which requires all branches to
1426     match the same length of string. An assertion such as
1427     <pre>
1428     (?&#60;=ab(c|de))
1429 nigel 75 </pre>
1430 nigel 63 is not permitted, because its single top-level branch can match two different
1431     lengths, but it is acceptable if rewritten to use two top-level branches:
1432     <pre>
1433     (?&#60;=abc|abde)
1434 nigel 75 </pre>
1435 nigel 63 The implementation of lookbehind assertions is, for each alternative, to
1436 nigel 93 temporarily move the current position back by the fixed length and then try to
1437 nigel 63 match. If there are insufficient characters before the current position, the
1438 nigel 93 assertion fails.
1439 nigel 63 </P>
1440     <P>
1441     PCRE does not allow the \C escape (which matches a single byte in UTF-8 mode)
1442     to appear in lookbehind assertions, because it makes it impossible to calculate
1443 nigel 93 the length of the lookbehind. The \X and \R escapes, which can match
1444     different numbers of bytes, are also not permitted.
1445 nigel 63 </P>
1446     <P>
1447 nigel 93 Possessive quantifiers can be used in conjunction with lookbehind assertions to
1448     specify efficient matching at the end of the subject string. Consider a simple
1449     pattern such as
1450 nigel 63 <pre>
1451     abcd$
1452 nigel 75 </pre>
1453 nigel 63 when applied to a long string that does not match. Because matching proceeds
1454     from left to right, PCRE will look for each "a" in the subject and then see if
1455     what follows matches the rest of the pattern. If the pattern is specified as
1456     <pre>
1457     ^.*abcd$
1458 nigel 75 </pre>
1459 nigel 63 the initial .* matches the entire string at first, but when this fails (because
1460     there is no following "a"), it backtracks to match all but the last character,
1461     then all but the last two characters, and so on. Once again the search for "a"
1462     covers the entire string, from right to left, so we are no better off. However,
1463     if the pattern is written as
1464     <pre>
1465     ^.*+(?&#60;=abcd)
1466 nigel 75 </pre>
1467 nigel 93 there can be no backtracking for the .*+ item; it can match only the entire
1468 nigel 63 string. The subsequent lookbehind assertion does a single test on the last four
1469     characters. If it fails, the match fails immediately. For long strings, this
1470     approach makes a significant difference to the processing time.
1471     </P>
1472 nigel 75 <br><b>
1473     Using multiple assertions
1474     </b><br>
1475 nigel 63 <P>
1476     Several assertions (of any sort) may occur in succession. For example,
1477     <pre>
1478     (?&#60;=\d{3})(?&#60;!999)foo
1479 nigel 75 </pre>
1480 nigel 63 matches "foo" preceded by three digits that are not "999". Notice that each of
1481     the assertions is applied independently at the same point in the subject
1482     string. First there is a check that the previous three characters are all
1483     digits, and then there is a check that the same three characters are not "999".
1484     This pattern does <i>not</i> match "foo" preceded by six characters, the first
1485     of which are digits and the last three of which are not "999". For example, it
1486     doesn't match "123abcfoo". A pattern to do that is
1487     <pre>
1488     (?&#60;=\d{3}...)(?&#60;!999)foo
1489 nigel 75 </pre>
1490 nigel 63 This time the first assertion looks at the preceding six characters, checking
1491     that the first three are digits, and then the second assertion checks that the
1492     preceding three characters are not "999".
1493     </P>
1494     <P>
1495     Assertions can be nested in any combination. For example,
1496     <pre>
1497     (?&#60;=(?&#60;!foo)bar)baz
1498 nigel 75 </pre>
1499 nigel 63 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1500     preceded by "foo", while
1501     <pre>
1502     (?&#60;=\d{3}(?!999)...)foo
1503 nigel 75 </pre>
1504     is another pattern that matches "foo" preceded by three digits and any three
1505 nigel 63 characters that are not "999".
1506 nigel 91 <a name="conditions"></a></P>
1507 nigel 93 <br><a name="SEC17" href="#TOC1">CONDITIONAL SUBPATTERNS</a><br>
1508 nigel 63 <P>
1509     It is possible to cause the matching process to obey a subpattern
1510     conditionally or to choose between two alternative subpatterns, depending on
1511     the result of an assertion, or whether a previous capturing subpattern matched
1512     or not. The two possible forms of conditional subpattern are
1513     <pre>
1514     (?(condition)yes-pattern)
1515     (?(condition)yes-pattern|no-pattern)
1516 nigel 75 </pre>
1517 nigel 63 If the condition is satisfied, the yes-pattern is used; otherwise the
1518     no-pattern (if present) is used. If there are more than two alternatives in the
1519     subpattern, a compile-time error occurs.
1520     </P>
1521     <P>
1522 nigel 93 There are four kinds of condition: references to subpatterns, references to
1523     recursion, a pseudo-condition called DEFINE, and assertions.
1524 nigel 91 </P>
1525 nigel 93 <br><b>
1526     Checking for a used subpattern by number
1527     </b><br>
1528 nigel 91 <P>
1529 nigel 93 If the text between the parentheses consists of a sequence of digits, the
1530     condition is true if the capturing subpattern of that number has previously
1531     matched.
1532     </P>
1533     <P>
1534 nigel 91 Consider the following pattern, which contains non-significant white space to
1535     make it more readable (assume the PCRE_EXTENDED option) and to divide it into
1536     three parts for ease of discussion:
1537 nigel 63 <pre>
1538     ( \( )? [^()]+ (?(1) \) )
1539 nigel 75 </pre>
1540 nigel 63 The first part matches an optional opening parenthesis, and if that
1541     character is present, sets it as the first captured substring. The second part
1542     matches one or more characters that are not parentheses. The third part is a
1543     conditional subpattern that tests whether the first set of parentheses matched
1544     or not. If they did, that is, if subject started with an opening parenthesis,
1545     the condition is true, and so the yes-pattern is executed and a closing
1546     parenthesis is required. Otherwise, since no-pattern is not present, the
1547     subpattern matches nothing. In other words, this pattern matches a sequence of
1548 nigel 93 non-parentheses, optionally enclosed in parentheses.
1549     </P>
1550     <br><b>
1551     Checking for a used subpattern by name
1552     </b><br>
1553     <P>
1554     Perl uses the syntax (?(&#60;name&#62;)...) or (?('name')...) to test for a used
1555     subpattern by name. For compatibility with earlier versions of PCRE, which had
1556     this facility before Perl, the syntax (?(name)...) is also recognized. However,
1557     there is a possible ambiguity with this syntax, because subpattern names may
1558     consist entirely of digits. PCRE looks first for a named subpattern; if it
1559     cannot find one and the name consists entirely of digits, PCRE looks for a
1560     subpattern of that number, which must be greater than zero. Using subpattern
1561     names that consist entirely of digits is not recommended.
1562     </P>
1563     <P>
1564     Rewriting the above example to use a named subpattern gives this:
1565 nigel 91 <pre>
1566 nigel 93 (?&#60;OPEN&#62; \( )? [^()]+ (?(&#60;OPEN&#62;) \) )
1567    
1568     </PRE>
1569     </P>
1570     <br><b>
1571     Checking for pattern recursion
1572     </b><br>
1573     <P>
1574     If the condition is the string (R), and there is no subpattern with the name R,
1575     the condition is true if a recursive call to the whole pattern or any
1576     subpattern has been made. If digits or a name preceded by ampersand follow the
1577     letter R, for example:
1578     <pre>
1579     (?(R3)...) or (?(R&name)...)
1580 nigel 91 </pre>
1581 nigel 93 the condition is true if the most recent recursion is into the subpattern whose
1582     number or name is given. This condition does not check the entire recursion
1583     stack.
1584 nigel 63 </P>
1585     <P>
1586 nigel 93 At "top level", all these recursion test conditions are false. Recursive
1587     patterns are described below.
1588     </P>
1589     <br><b>
1590     Defining subpatterns for use by reference only
1591     </b><br>
1592     <P>
1593     If the condition is the string (DEFINE), and there is no subpattern with the
1594     name DEFINE, the condition is always false. In this case, there may be only one
1595     alternative in the subpattern. It is always skipped if control reaches this
1596     point in the pattern; the idea of DEFINE is that it can be used to define
1597     "subroutines" that can be referenced from elsewhere. (The use of "subroutines"
1598     is described below.) For example, a pattern to match an IPv4 address could be
1599     written like this (ignore whitespace and line breaks):
1600     <pre>
1601     (?(DEFINE) (?&#60;byte&#62; 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
1602     \b (?&byte) (\.(?&byte)){3} \b
1603     </pre>
1604     The first part of the pattern is a DEFINE group inside which a another group
1605     named "byte" is defined. This matches an individual component of an IPv4
1606     address (a number less than 256). When matching takes place, this part of the
1607     pattern is skipped because DEFINE acts like a false condition.
1608     </P>
1609     <P>
1610     The rest of the pattern uses references to the named group to match the four
1611     dot-separated components of an IPv4 address, insisting on a word boundary at
1612     each end.
1613     </P>
1614     <br><b>
1615     Assertion conditions
1616     </b><br>
1617     <P>
1618     If the condition is not in any of the above formats, it must be an assertion.
1619 nigel 63 This may be a positive or negative lookahead or lookbehind assertion. Consider
1620     this pattern, again containing non-significant white space, and with the two
1621     alternatives on the second line:
1622     <pre>
1623     (?(?=[^a-z]*[a-z])
1624     \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
1625 nigel 75 </pre>
1626 nigel 63 The condition is a positive lookahead assertion that matches an optional
1627     sequence of non-letters followed by a letter. In other words, it tests for the
1628     presence of at least one letter in the subject. If a letter is found, the
1629     subject is matched against the first alternative; otherwise it is matched
1630     against the second. This pattern matches strings in one of the two forms
1631     dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
1632 nigel 75 <a name="comments"></a></P>
1633 nigel 93 <br><a name="SEC18" href="#TOC1">COMMENTS</a><br>
1634 nigel 63 <P>
1635 nigel 75 The sequence (?# marks the start of a comment that continues up to the next
1636 nigel 63 closing parenthesis. Nested parentheses are not permitted. The characters
1637     that make up a comment play no part in the pattern matching at all.
1638     </P>
1639     <P>
1640     If the PCRE_EXTENDED option is set, an unescaped # character outside a
1641 nigel 91 character class introduces a comment that continues to immediately after the
1642     next newline in the pattern.
1643     <a name="recursion"></a></P>
1644 nigel 93 <br><a name="SEC19" href="#TOC1">RECURSIVE PATTERNS</a><br>
1645 nigel 63 <P>
1646     Consider the problem of matching a string in parentheses, allowing for
1647     unlimited nested parentheses. Without the use of recursion, the best that can
1648     be done is to use a pattern that matches up to some fixed depth of nesting. It
1649 nigel 93 is not possible to handle an arbitrary nesting depth.
1650     </P>
1651     <P>
1652     For some time, Perl has provided a facility that allows regular expressions to
1653     recurse (amongst other things). It does this by interpolating Perl code in the
1654     expression at run time, and the code can refer to the expression itself. A Perl
1655     pattern using code interpolation to solve the parentheses problem can be
1656     created like this:
1657 nigel 63 <pre>
1658     $re = qr{\( (?: (?&#62;[^()]+) | (?p{$re}) )* \)}x;
1659 nigel 75 </pre>
1660 nigel 63 The (?p{...}) item interpolates Perl code at run time, and in this case refers
1661 nigel 93 recursively to the pattern in which it appears.
1662 nigel 63 </P>
1663     <P>
1664 nigel 93 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
1665     supports special syntax for recursion of the entire pattern, and also for
1666     individual subpattern recursion. After its introduction in PCRE and Python,
1667     this kind of recursion was introduced into Perl at release 5.10.
1668 nigel 63 </P>
1669     <P>
1670 nigel 93 A special item that consists of (? followed by a number greater than zero and a
1671     closing parenthesis is a recursive call of the subpattern of the given number,
1672     provided that it occurs inside that subpattern. (If not, it is a "subroutine"
1673     call, which is described in the next section.) The special item (?R) or (?0) is
1674     a recursive call of the entire regular expression.
1675 nigel 87 </P>
1676     <P>
1677 nigel 93 In PCRE (like Python, but unlike Perl), a recursive subpattern call is always
1678     treated as an atomic group. That is, once it has matched some of the subject
1679     string, it is never re-entered, even if it contains untried alternatives and
1680     there is a subsequent matching failure.
1681     </P>
1682     <P>
1683 nigel 87 This PCRE pattern solves the nested parentheses problem (assume the
1684     PCRE_EXTENDED option is set so that white space is ignored):
1685 nigel 63 <pre>
1686     \( ( (?&#62;[^()]+) | (?R) )* \)
1687 nigel 75 </pre>
1688 nigel 63 First it matches an opening parenthesis. Then it matches any number of
1689     substrings which can either be a sequence of non-parentheses, or a recursive
1690 nigel 87 match of the pattern itself (that is, a correctly parenthesized substring).
1691 nigel 63 Finally there is a closing parenthesis.
1692     </P>
1693     <P>
1694     If this were part of a larger pattern, you would not want to recurse the entire
1695     pattern, so instead you could use this:
1696     <pre>
1697     ( \( ( (?&#62;[^()]+) | (?1) )* \) )
1698 nigel 75 </pre>
1699 nigel 63 We have put the pattern into parentheses, and caused the recursion to refer to
1700     them instead of the whole pattern. In a larger pattern, keeping track of
1701     parenthesis numbers can be tricky. It may be more convenient to use named
1702 nigel 93 parentheses instead. The Perl syntax for this is (?&name); PCRE's earlier
1703     syntax (?P&#62;name) is also supported. We could rewrite the above example as
1704     follows:
1705 nigel 63 <pre>
1706 nigel 93 (?&#60;pn&#62; \( ( (?&#62;[^()]+) | (?&pn) )* \) )
1707 nigel 75 </pre>
1708 nigel 93 If there is more than one subpattern with the same name, the earliest one is
1709     used. This particular example pattern contains nested unlimited repeats, and so
1710     the use of atomic grouping for matching strings of non-parentheses is important
1711 nigel 63 when applying the pattern to strings that do not match. For example, when this
1712     pattern is applied to
1713     <pre>
1714     (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
1715 nigel 75 </pre>
1716 nigel 63 it yields "no match" quickly. However, if atomic grouping is not used,
1717     the match runs for a very long time indeed because there are so many different
1718     ways the + and * repeats can carve up the subject, and all have to be tested
1719     before failure can be reported.
1720     </P>
1721     <P>
1722     At the end of a match, the values set for any capturing subpatterns are those
1723     from the outermost level of the recursion at which the subpattern value is set.
1724     If you want to obtain intermediate values, a callout function can be used (see
1725 nigel 93 below and the
1726 nigel 63 <a href="pcrecallout.html"><b>pcrecallout</b></a>
1727     documentation). If the pattern above is matched against
1728     <pre>
1729     (ab(cd)ef)
1730 nigel 75 </pre>
1731 nigel 63 the value for the capturing parentheses is "ef", which is the last value taken
1732     on at the top level. If additional parentheses are added, giving
1733     <pre>
1734     \( ( ( (?&#62;[^()]+) | (?R) )* ) \)
1735     ^ ^
1736     ^ ^
1737 nigel 75 </pre>
1738 nigel 63 the string they capture is "ab(cd)ef", the contents of the top level
1739     parentheses. If there are more than 15 capturing parentheses in a pattern, PCRE
1740     has to obtain extra memory to store data during a recursion, which it does by
1741     using <b>pcre_malloc</b>, freeing it via <b>pcre_free</b> afterwards. If no
1742     memory can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
1743     </P>
1744     <P>
1745     Do not confuse the (?R) item with the condition (R), which tests for recursion.
1746     Consider this pattern, which matches text in angle brackets, allowing for
1747     arbitrary nesting. Only digits are allowed in nested brackets (that is, when
1748     recursing), whereas any characters are permitted at the outer level.
1749     <pre>
1750     &#60; (?: (?(R) \d++ | [^&#60;&#62;]*+) | (?R)) * &#62;
1751 nigel 75 </pre>
1752 nigel 63 In this pattern, (?(R) is the start of a conditional subpattern, with two
1753     different alternatives for the recursive and non-recursive cases. The (?R) item
1754     is the actual recursive call.
1755 nigel 75 <a name="subpatternsassubroutines"></a></P>
1756 nigel 93 <br><a name="SEC20" href="#TOC1">SUBPATTERNS AS SUBROUTINES</a><br>
1757 nigel 63 <P>
1758     If the syntax for a recursive subpattern reference (either by number or by
1759     name) is used outside the parentheses to which it refers, it operates like a
1760 nigel 93 subroutine in a programming language. The "called" subpattern may be defined
1761     before or after the reference. An earlier example pointed out that the pattern
1762 nigel 63 <pre>
1763     (sens|respons)e and \1ibility
1764 nigel 75 </pre>
1765 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1766     "sense and responsibility". If instead the pattern
1767     <pre>
1768     (sens|respons)e and (?1)ibility
1769 nigel 75 </pre>
1770 nigel 63 is used, it does match "sense and responsibility" as well as the other two
1771 nigel 93 strings. Another example is given in the discussion of DEFINE above.
1772 nigel 63 </P>
1773 nigel 87 <P>
1774     Like recursive subpatterns, a "subroutine" call is always treated as an atomic
1775     group. That is, once it has matched some of the subject string, it is never
1776     re-entered, even if it contains untried alternatives and there is a subsequent
1777     matching failure.
1778     </P>
1779 nigel 63 <P>
1780 nigel 93 When a subpattern is used as a subroutine, processing options such as
1781     case-independence are fixed when the subpattern is defined. They cannot be
1782     changed for different calls. For example, consider this pattern:
1783     <pre>
1784     (abc)(?i:(?1))
1785     </pre>
1786     It matches "abcabc". It does not match "abcABC" because the change of
1787     processing option does not affect the called subpattern.
1788     </P>
1789     <br><a name="SEC21" href="#TOC1">CALLOUTS</a><br>
1790     <P>
1791 nigel 63 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
1792     code to be obeyed in the middle of matching a regular expression. This makes it
1793     possible, amongst other things, to extract different substrings that match the
1794     same pair of parentheses when there is a repetition.
1795     </P>
1796     <P>
1797     PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
1798     code. The feature is called "callout". The caller of PCRE provides an external
1799     function by putting its entry point in the global variable <i>pcre_callout</i>.
1800     By default, this variable contains NULL, which disables all calling out.
1801     </P>
1802     <P>
1803     Within a regular expression, (?C) indicates the points at which the external
1804     function is to be called. If you want to identify different callout points, you
1805     can put a number less than 256 after the letter C. The default value is zero.
1806     For example, this pattern has two callout points:
1807     <pre>
1808     (?C1)\dabc(?C2)def
1809 nigel 75 </pre>
1810     If the PCRE_AUTO_CALLOUT flag is passed to <b>pcre_compile()</b>, callouts are
1811     automatically installed before each item in the pattern. They are all numbered
1812     255.
1813 nigel 63 </P>
1814     <P>
1815     During matching, when PCRE reaches a callout point (and <i>pcre_callout</i> is
1816     set), the external function is called. It is provided with the number of the
1817 nigel 75 callout, the position in the pattern, and, optionally, one item of data
1818     originally supplied by the caller of <b>pcre_exec()</b>. The callout function
1819     may cause matching to proceed, to backtrack, or to fail altogether. A complete
1820     description of the interface to the callout function is given in the
1821 nigel 63 <a href="pcrecallout.html"><b>pcrecallout</b></a>
1822     documentation.
1823     </P>
1824 nigel 93 <br><a name="SEC22" href="#TOC1">SEE ALSO</a><br>
1825 nigel 63 <P>
1826 nigel 93 <b>pcreapi</b>(3), <b>pcrecallout</b>(3), <b>pcrematching</b>(3), <b>pcre</b>(3).
1827     </P>
1828 ph10 99 <br><a name="SEC23" href="#TOC1">AUTHOR</a><br>
1829 nigel 93 <P>
1830 ph10 99 Philip Hazel
1831 nigel 63 <br>
1832 ph10 99 University Computing Service
1833     <br>
1834     Cambridge CB2 3QH, England.
1835     <br>
1836     </P>
1837     <br><a name="SEC24" href="#TOC1">REVISION</a><br>
1838     <P>
1839     Last updated: 06 March 2007
1840     <br>
1841     Copyright &copy; 1997-2007 University of Cambridge.
1842     <br>
1843 nigel 75 <p>
1844     Return to the <a href="index.html">PCRE index page</a>.
1845     </p>

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12