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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 87 - (hide annotations) (download)
Sat Feb 24 21:41:21 2007 UTC (7 years, 1 month ago) by nigel
File size: 65123 byte(s)
Load pcre-6.5 into code/trunk.

1 nigel 79 .TH PCREPATTERN 3
2 nigel 63 .SH NAME
3     PCRE - Perl-compatible regular expressions
4 nigel 75 .SH "PCRE REGULAR EXPRESSION DETAILS"
5 nigel 63 .rs
6     .sp
7     The syntax and semantics of the regular expressions supported by PCRE are
8     described below. Regular expressions are also described in the Perl
9 nigel 75 documentation and in a number of books, some of which have copious examples.
10     Jeffrey Friedl's "Mastering Regular Expressions", published by O'Reilly, covers
11     regular expressions in great detail. This description of PCRE's regular
12     expressions is intended as reference material.
13     .P
14     The original operation of PCRE was on strings of one-byte characters. However,
15     there is now also support for UTF-8 character strings. To use this, you must
16     build PCRE to include UTF-8 support, and then call \fBpcre_compile()\fP with
17     the PCRE_UTF8 option. How this affects pattern matching is mentioned in several
18     places below. There is also a summary of UTF-8 features in the
19 nigel 63 .\" HTML <a href="pcre.html#utf8support">
20     .\" </a>
21     section on UTF-8 support
22     .\"
23     in the main
24     .\" HREF
25 nigel 75 \fBpcre\fP
26 nigel 63 .\"
27     page.
28 nigel 75 .P
29 nigel 77 The remainder of this document discusses the patterns that are supported by
30     PCRE when its main matching function, \fBpcre_exec()\fP, is used.
31     From release 6.0, PCRE offers a second matching function,
32     \fBpcre_dfa_exec()\fP, which matches using a different algorithm that is not
33     Perl-compatible. The advantages and disadvantages of the alternative function,
34     and how it differs from the normal function, are discussed in the
35     .\" HREF
36     \fBpcrematching\fP
37     .\"
38     page.
39     .P
40 nigel 63 A regular expression is a pattern that is matched against a subject string from
41     left to right. Most characters stand for themselves in a pattern, and match the
42     corresponding characters in the subject. As a trivial example, the pattern
43 nigel 75 .sp
44 nigel 63 The quick brown fox
45 nigel 75 .sp
46 nigel 77 matches a portion of a subject string that is identical to itself. When
47     caseless matching is specified (the PCRE_CASELESS option), letters are matched
48     independently of case. In UTF-8 mode, PCRE always understands the concept of
49     case for characters whose values are less than 128, so caseless matching is
50     always possible. For characters with higher values, the concept of case is
51     supported if PCRE is compiled with Unicode property support, but not otherwise.
52     If you want to use caseless matching for characters 128 and above, you must
53     ensure that PCRE is compiled with Unicode property support as well as with
54     UTF-8 support.
55     .P
56     The power of regular expressions comes from the ability to include alternatives
57     and repetitions in the pattern. These are encoded in the pattern by the use of
58 nigel 75 \fImetacharacters\fP, which do not stand for themselves but instead are
59 nigel 63 interpreted in some special way.
60 nigel 75 .P
61     There are two different sets of metacharacters: those that are recognized
62 nigel 63 anywhere in the pattern except within square brackets, and those that are
63 nigel 75 recognized in square brackets. Outside square brackets, the metacharacters are
64 nigel 63 as follows:
65 nigel 75 .sp
66     \e general escape character with several uses
67 nigel 63 ^ assert start of string (or line, in multiline mode)
68     $ assert end of string (or line, in multiline mode)
69     . match any character except newline (by default)
70     [ start character class definition
71     | start of alternative branch
72     ( start subpattern
73     ) end subpattern
74     ? extends the meaning of (
75     also 0 or 1 quantifier
76     also quantifier minimizer
77     * 0 or more quantifier
78     + 1 or more quantifier
79     also "possessive quantifier"
80     { start min/max quantifier
81 nigel 75 .sp
82 nigel 63 Part of a pattern that is in square brackets is called a "character class". In
83 nigel 75 a character class the only metacharacters are:
84     .sp
85     \e general escape character
86 nigel 63 ^ negate the class, but only if the first character
87     - indicates character range
88 nigel 75 .\" JOIN
89 nigel 63 [ POSIX character class (only if followed by POSIX
90     syntax)
91     ] terminates the character class
92 nigel 75 .sp
93     The following sections describe the use of each of the metacharacters.
94     .
95 nigel 63 .SH BACKSLASH
96     .rs
97     .sp
98     The backslash character has several uses. Firstly, if it is followed by a
99 nigel 75 non-alphanumeric character, it takes away any special meaning that character may
100 nigel 63 have. This use of backslash as an escape character applies both inside and
101     outside character classes.
102 nigel 75 .P
103     For example, if you want to match a * character, you write \e* in the pattern.
104 nigel 63 This escaping action applies whether or not the following character would
105 nigel 75 otherwise be interpreted as a metacharacter, so it is always safe to precede a
106     non-alphanumeric with backslash to specify that it stands for itself. In
107     particular, if you want to match a backslash, you write \e\e.
108     .P
109 nigel 63 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
110     pattern (other than in a character class) and characters between a # outside
111     a character class and the next newline character are ignored. An escaping
112     backslash can be used to include a whitespace or # character as part of the
113     pattern.
114 nigel 75 .P
115 nigel 63 If you want to remove the special meaning from a sequence of characters, you
116 nigel 75 can do so by putting them between \eQ and \eE. This is different from Perl in
117     that $ and @ are handled as literals in \eQ...\eE sequences in PCRE, whereas in
118 nigel 63 Perl, $ and @ cause variable interpolation. Note the following examples:
119 nigel 75 .sp
120 nigel 63 Pattern PCRE matches Perl matches
121 nigel 75 .sp
122     .\" JOIN
123     \eQabc$xyz\eE abc$xyz abc followed by the
124 nigel 63 contents of $xyz
125 nigel 75 \eQabc\e$xyz\eE abc\e$xyz abc\e$xyz
126     \eQabc\eE\e$\eQxyz\eE abc$xyz abc$xyz
127     .sp
128     The \eQ...\eE sequence is recognized both inside and outside character classes.
129     .
130     .
131     .\" HTML <a name="digitsafterbackslash"></a>
132     .SS "Non-printing characters"
133     .rs
134     .sp
135 nigel 63 A second use of backslash provides a way of encoding non-printing characters
136     in patterns in a visible manner. There is no restriction on the appearance of
137     non-printing characters, apart from the binary zero that terminates a pattern,
138     but when a pattern is being prepared by text editing, it is usually easier to
139     use one of the following escape sequences than the binary character it
140     represents:
141 nigel 75 .sp
142     \ea alarm, that is, the BEL character (hex 07)
143     \ecx "control-x", where x is any character
144     \ee escape (hex 1B)
145     \ef formfeed (hex 0C)
146     \en newline (hex 0A)
147     \er carriage return (hex 0D)
148     \et tab (hex 09)
149     \eddd character with octal code ddd, or backreference
150     \exhh character with hex code hh
151 nigel 87 \ex{hhh..} character with hex code hhh..
152 nigel 75 .sp
153     The precise effect of \ecx is as follows: if x is a lower case letter, it
154 nigel 63 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
155 nigel 75 Thus \ecz becomes hex 1A, but \ec{ becomes hex 3B, while \ec; becomes hex
156 nigel 63 7B.
157 nigel 75 .P
158     After \ex, from zero to two hexadecimal digits are read (letters can be in
159 nigel 87 upper or lower case). Any number of hexadecimal digits may appear between \ex{
160     and }, but the value of the character code must be less than 256 in non-UTF-8
161     mode, and less than 2**31 in UTF-8 mode (that is, the maximum hexadecimal value
162     is 7FFFFFFF). If characters other than hexadecimal digits appear between \ex{
163     and }, or if there is no terminating }, this form of escape is not recognized.
164     Instead, the initial \ex will be interpreted as a basic hexadecimal escape,
165     with no following digits, giving a character whose value is zero.
166 nigel 75 .P
167 nigel 63 Characters whose value is less than 256 can be defined by either of the two
168 nigel 87 syntaxes for \ex. There is no difference in the way they are handled. For
169     example, \exdc is exactly the same as \ex{dc}.
170 nigel 75 .P
171     After \e0 up to two further octal digits are read. In both cases, if there
172 nigel 63 are fewer than two digits, just those that are present are used. Thus the
173 nigel 75 sequence \e0\ex\e07 specifies two binary zeros followed by a BEL character
174 nigel 63 (code value 7). Make sure you supply two digits after the initial zero if the
175 nigel 75 pattern character that follows is itself an octal digit.
176     .P
177 nigel 63 The handling of a backslash followed by a digit other than 0 is complicated.
178     Outside a character class, PCRE reads it and any following digits as a decimal
179     number. If the number is less than 10, or if there have been at least that many
180     previous capturing left parentheses in the expression, the entire sequence is
181 nigel 75 taken as a \fIback reference\fP. A description of how this works is given
182     .\" HTML <a href="#backreferences">
183     .\" </a>
184     later,
185     .\"
186     following the discussion of
187     .\" HTML <a href="#subpattern">
188     .\" </a>
189     parenthesized subpatterns.
190     .\"
191     .P
192 nigel 63 Inside a character class, or if the decimal number is greater than 9 and there
193     have not been that many capturing subpatterns, PCRE re-reads up to three octal
194     digits following the backslash, and generates a single byte from the least
195     significant 8 bits of the value. Any subsequent digits stand for themselves.
196     For example:
197 nigel 75 .sp
198     \e040 is another way of writing a space
199     .\" JOIN
200     \e40 is the same, provided there are fewer than 40
201 nigel 63 previous capturing subpatterns
202 nigel 75 \e7 is always a back reference
203     .\" JOIN
204     \e11 might be a back reference, or another way of
205 nigel 63 writing a tab
206 nigel 75 \e011 is always a tab
207     \e0113 is a tab followed by the character "3"
208     .\" JOIN
209     \e113 might be a back reference, otherwise the
210 nigel 63 character with octal code 113
211 nigel 75 .\" JOIN
212     \e377 might be a back reference, otherwise
213 nigel 63 the byte consisting entirely of 1 bits
214 nigel 75 .\" JOIN
215     \e81 is either a back reference, or a binary zero
216 nigel 63 followed by the two characters "8" and "1"
217 nigel 75 .sp
218 nigel 63 Note that octal values of 100 or greater must not be introduced by a leading
219     zero, because no more than three octal digits are ever read.
220 nigel 75 .P
221 nigel 63 All the sequences that define a single byte value or a single UTF-8 character
222     (in UTF-8 mode) can be used both inside and outside character classes. In
223 nigel 75 addition, inside a character class, the sequence \eb is interpreted as the
224     backspace character (hex 08), and the sequence \eX is interpreted as the
225     character "X". Outside a character class, these sequences have different
226     meanings
227     .\" HTML <a href="#uniextseq">
228     .\" </a>
229     (see below).
230     .\"
231     .
232     .
233     .SS "Generic character types"
234     .rs
235     .sp
236     The third use of backslash is for specifying generic character types. The
237     following are always recognized:
238     .sp
239     \ed any decimal digit
240     \eD any character that is not a decimal digit
241     \es any whitespace character
242     \eS any character that is not a whitespace character
243     \ew any "word" character
244     \eW any "non-word" character
245     .sp
246 nigel 63 Each pair of escape sequences partitions the complete set of characters into
247     two disjoint sets. Any given character matches one, and only one, of each pair.
248 nigel 75 .P
249     These character type sequences can appear both inside and outside character
250     classes. They each match one character of the appropriate type. If the current
251     matching point is at the end of the subject string, all of them fail, since
252     there is no character to match.
253     .P
254     For compatibility with Perl, \es does not match the VT character (code 11).
255     This makes it different from the the POSIX "space" class. The \es characters
256 nigel 63 are HT (9), LF (10), FF (12), CR (13), and space (32).
257 nigel 75 .P
258     A "word" character is an underscore or any character less than 256 that is a
259     letter or digit. The definition of letters and digits is controlled by PCRE's
260     low-valued character tables, and may vary if locale-specific matching is taking
261     place (see
262 nigel 63 .\" HTML <a href="pcreapi.html#localesupport">
263     .\" </a>
264     "Locale support"
265     .\"
266     in the
267     .\" HREF
268 nigel 75 \fBpcreapi\fP
269 nigel 63 .\"
270 nigel 75 page). For example, in the "fr_FR" (French) locale, some character codes
271     greater than 128 are used for accented letters, and these are matched by \ew.
272     .P
273     In UTF-8 mode, characters with values greater than 128 never match \ed, \es, or
274     \ew, and always match \eD, \eS, and \eW. This is true even when Unicode
275 nigel 87 character property support is available. The use of locales with Unicode is
276     discouraged.
277 nigel 75 .
278     .
279     .\" HTML <a name="uniextseq"></a>
280     .SS Unicode character properties
281     .rs
282     .sp
283     When PCRE is built with Unicode character property support, three additional
284 nigel 87 escape sequences to match character properties are available when UTF-8 mode
285 nigel 75 is selected. They are:
286     .sp
287 nigel 87 \ep{\fIxx\fP} a character with the \fIxx\fP property
288     \eP{\fIxx\fP} a character without the \fIxx\fP property
289     \eX an extended Unicode sequence
290 nigel 75 .sp
291 nigel 87 The property names represented by \fIxx\fP above are limited to the Unicode
292     script names, the general category properties, and "Any", which matches any
293     character (including newline). Other properties such as "InMusicalSymbols" are
294     not currently supported by PCRE. Note that \eP{Any} does not match any
295     characters, so always causes a match failure.
296 nigel 75 .P
297 nigel 87 Sets of Unicode characters are defined as belonging to certain scripts. A
298     character from one of these sets can be matched using a script name. For
299     example:
300 nigel 75 .sp
301 nigel 87 \ep{Greek}
302     \eP{Han}
303     .sp
304     Those that are not part of an identified script are lumped together as
305     "Common". The current list of scripts is:
306     .P
307     Arabic,
308     Armenian,
309     Bengali,
310     Bopomofo,
311     Braille,
312     Buginese,
313     Buhid,
314     Canadian_Aboriginal,
315     Cherokee,
316     Common,
317     Coptic,
318     Cypriot,
319     Cyrillic,
320     Deseret,
321     Devanagari,
322     Ethiopic,
323     Georgian,
324     Glagolitic,
325     Gothic,
326     Greek,
327     Gujarati,
328     Gurmukhi,
329     Han,
330     Hangul,
331     Hanunoo,
332     Hebrew,
333     Hiragana,
334     Inherited,
335     Kannada,
336     Katakana,
337     Kharoshthi,
338     Khmer,
339     Lao,
340     Latin,
341     Limbu,
342     Linear_B,
343     Malayalam,
344     Mongolian,
345     Myanmar,
346     New_Tai_Lue,
347     Ogham,
348     Old_Italic,
349     Old_Persian,
350     Oriya,
351     Osmanya,
352     Runic,
353     Shavian,
354     Sinhala,
355     Syloti_Nagri,
356     Syriac,
357     Tagalog,
358     Tagbanwa,
359     Tai_Le,
360     Tamil,
361     Telugu,
362     Thaana,
363     Thai,
364     Tibetan,
365     Tifinagh,
366     Ugaritic,
367     Yi.
368     .P
369     Each character has exactly one general category property, specified by a
370     two-letter abbreviation. For compatibility with Perl, negation can be specified
371     by including a circumflex between the opening brace and the property name. For
372     example, \ep{^Lu} is the same as \eP{Lu}.
373     .P
374     If only one letter is specified with \ep or \eP, it includes all the general
375     category properties that start with that letter. In this case, in the absence
376     of negation, the curly brackets in the escape sequence are optional; these two
377     examples have the same effect:
378     .sp
379 nigel 75 \ep{L}
380     \epL
381     .sp
382 nigel 87 The following general category property codes are supported:
383 nigel 75 .sp
384     C Other
385     Cc Control
386     Cf Format
387     Cn Unassigned
388     Co Private use
389     Cs Surrogate
390     .sp
391     L Letter
392     Ll Lower case letter
393     Lm Modifier letter
394     Lo Other letter
395     Lt Title case letter
396     Lu Upper case letter
397     .sp
398     M Mark
399     Mc Spacing mark
400     Me Enclosing mark
401     Mn Non-spacing mark
402     .sp
403     N Number
404     Nd Decimal number
405     Nl Letter number
406     No Other number
407     .sp
408     P Punctuation
409     Pc Connector punctuation
410     Pd Dash punctuation
411     Pe Close punctuation
412     Pf Final punctuation
413     Pi Initial punctuation
414     Po Other punctuation
415     Ps Open punctuation
416     .sp
417     S Symbol
418     Sc Currency symbol
419     Sk Modifier symbol
420     Sm Mathematical symbol
421     So Other symbol
422     .sp
423     Z Separator
424     Zl Line separator
425     Zp Paragraph separator
426     Zs Space separator
427     .sp
428 nigel 87 The special property L& is also supported: it matches a character that has
429     the Lu, Ll, or Lt property, in other words, a letter that is not classified as
430     a modifier or "other".
431 nigel 75 .P
432 nigel 87 The long synonyms for these properties that Perl supports (such as \ep{Letter})
433     are not supported by PCRE. Nor is is permitted to prefix any of these
434     properties with "Is".
435     .P
436     No character that is in the Unicode table has the Cn (unassigned) property.
437     Instead, this property is assumed for any code point that is not in the
438     Unicode table.
439     .P
440 nigel 75 Specifying caseless matching does not affect these escape sequences. For
441     example, \ep{Lu} always matches only upper case letters.
442     .P
443     The \eX escape matches any number of Unicode characters that form an extended
444     Unicode sequence. \eX is equivalent to
445     .sp
446     (?>\ePM\epM*)
447     .sp
448     That is, it matches a character without the "mark" property, followed by zero
449     or more characters with the "mark" property, and treats the sequence as an
450     atomic group
451     .\" HTML <a href="#atomicgroup">
452     .\" </a>
453     (see below).
454     .\"
455     Characters with the "mark" property are typically accents that affect the
456     preceding character.
457     .P
458     Matching characters by Unicode property is not fast, because PCRE has to search
459     a structure that contains data for over fifteen thousand characters. That is
460     why the traditional escape sequences such as \ed and \ew do not use Unicode
461     properties in PCRE.
462     .
463     .
464     .\" HTML <a name="smallassertions"></a>
465     .SS "Simple assertions"
466     .rs
467     .sp
468 nigel 63 The fourth use of backslash is for certain simple assertions. An assertion
469     specifies a condition that has to be met at a particular point in a match,
470     without consuming any characters from the subject string. The use of
471 nigel 75 subpatterns for more complicated assertions is described
472     .\" HTML <a href="#bigassertions">
473     .\" </a>
474     below.
475     .\"
476     The backslashed
477     assertions are:
478     .sp
479     \eb matches at a word boundary
480     \eB matches when not at a word boundary
481     \eA matches at start of subject
482     \eZ matches at end of subject or before newline at end
483     \ez matches at end of subject
484     \eG matches at first matching position in subject
485     .sp
486     These assertions may not appear in character classes (but note that \eb has a
487 nigel 63 different meaning, namely the backspace character, inside a character class).
488 nigel 75 .P
489 nigel 63 A word boundary is a position in the subject string where the current character
490 nigel 75 and the previous character do not both match \ew or \eW (i.e. one matches
491     \ew and the other matches \eW), or the start or end of the string if the
492     first or last character matches \ew, respectively.
493     .P
494     The \eA, \eZ, and \ez assertions differ from the traditional circumflex and
495     dollar (described in the next section) in that they only ever match at the very
496     start and end of the subject string, whatever options are set. Thus, they are
497     independent of multiline mode. These three assertions are not affected by the
498     PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
499     circumflex and dollar metacharacters. However, if the \fIstartoffset\fP
500     argument of \fBpcre_exec()\fP is non-zero, indicating that matching is to start
501     at a point other than the beginning of the subject, \eA can never match. The
502     difference between \eZ and \ez is that \eZ matches before a newline that is the
503     last character of the string as well as at the end of the string, whereas \ez
504     matches only at the end.
505     .P
506     The \eG assertion is true only when the current matching position is at the
507     start point of the match, as specified by the \fIstartoffset\fP argument of
508     \fBpcre_exec()\fP. It differs from \eA when the value of \fIstartoffset\fP is
509     non-zero. By calling \fBpcre_exec()\fP multiple times with appropriate
510 nigel 63 arguments, you can mimic Perl's /g option, and it is in this kind of
511 nigel 75 implementation where \eG can be useful.
512     .P
513     Note, however, that PCRE's interpretation of \eG, as the start of the current
514 nigel 63 match, is subtly different from Perl's, which defines it as the end of the
515     previous match. In Perl, these can be different when the previously matched
516     string was empty. Because PCRE does just one match at a time, it cannot
517     reproduce this behaviour.
518 nigel 75 .P
519     If all the alternatives of a pattern begin with \eG, the expression is anchored
520 nigel 63 to the starting match position, and the "anchored" flag is set in the compiled
521     regular expression.
522 nigel 75 .
523     .
524     .SH "CIRCUMFLEX AND DOLLAR"
525 nigel 63 .rs
526     .sp
527     Outside a character class, in the default matching mode, the circumflex
528 nigel 75 character is an assertion that is true only if the current matching point is
529     at the start of the subject string. If the \fIstartoffset\fP argument of
530     \fBpcre_exec()\fP is non-zero, circumflex can never match if the PCRE_MULTILINE
531 nigel 63 option is unset. Inside a character class, circumflex has an entirely different
532 nigel 75 meaning
533     .\" HTML <a href="#characterclass">
534     .\" </a>
535     (see below).
536     .\"
537     .P
538 nigel 63 Circumflex need not be the first character of the pattern if a number of
539     alternatives are involved, but it should be the first thing in each alternative
540     in which it appears if the pattern is ever to match that branch. If all
541     possible alternatives start with a circumflex, that is, if the pattern is
542     constrained to match only at the start of the subject, it is said to be an
543     "anchored" pattern. (There are also other constructs that can cause a pattern
544     to be anchored.)
545 nigel 75 .P
546     A dollar character is an assertion that is true only if the current matching
547 nigel 63 point is at the end of the subject string, or immediately before a newline
548     character that is the last character in the string (by default). Dollar need
549     not be the last character of the pattern if a number of alternatives are
550     involved, but it should be the last item in any branch in which it appears.
551     Dollar has no special meaning in a character class.
552 nigel 75 .P
553 nigel 63 The meaning of dollar can be changed so that it matches only at the very end of
554     the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
555 nigel 75 does not affect the \eZ assertion.
556     .P
557 nigel 63 The meanings of the circumflex and dollar characters are changed if the
558     PCRE_MULTILINE option is set. When this is the case, they match immediately
559     after and immediately before an internal newline character, respectively, in
560     addition to matching at the start and end of the subject string. For example,
561 nigel 75 the pattern /^abc$/ matches the subject string "def\enabc" (where \en
562     represents a newline character) in multiline mode, but not otherwise.
563     Consequently, patterns that are anchored in single line mode because all
564     branches start with ^ are not anchored in multiline mode, and a match for
565     circumflex is possible when the \fIstartoffset\fP argument of \fBpcre_exec()\fP
566     is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is
567     set.
568     .P
569     Note that the sequences \eA, \eZ, and \ez can be used to match the start and
570 nigel 63 end of the subject in both modes, and if all branches of a pattern start with
571 nigel 75 \eA it is always anchored, whether PCRE_MULTILINE is set or not.
572     .
573     .
574     .SH "FULL STOP (PERIOD, DOT)"
575 nigel 63 .rs
576     .sp
577     Outside a character class, a dot in the pattern matches any one character in
578     the subject, including a non-printing character, but not (by default) newline.
579     In UTF-8 mode, a dot matches any UTF-8 character, which might be more than one
580 nigel 75 byte long, except (by default) newline. If the PCRE_DOTALL option is set,
581 nigel 63 dots match newlines as well. The handling of dot is entirely independent of the
582     handling of circumflex and dollar, the only relationship being that they both
583     involve newline characters. Dot has no special meaning in a character class.
584 nigel 75 .
585     .
586     .SH "MATCHING A SINGLE BYTE"
587 nigel 63 .rs
588     .sp
589 nigel 75 Outside a character class, the escape sequence \eC matches any one byte, both
590     in and out of UTF-8 mode. Unlike a dot, it can match a newline. The feature is
591     provided in Perl in order to match individual bytes in UTF-8 mode. Because it
592     breaks up UTF-8 characters into individual bytes, what remains in the string
593     may be a malformed UTF-8 string. For this reason, the \eC escape sequence is
594     best avoided.
595     .P
596     PCRE does not allow \eC to appear in lookbehind assertions
597     .\" HTML <a href="#lookbehind">
598     .\" </a>
599     (described below),
600     .\"
601     because in UTF-8 mode this would make it impossible to calculate the length of
602     the lookbehind.
603     .
604     .
605     .\" HTML <a name="characterclass"></a>
606     .SH "SQUARE BRACKETS AND CHARACTER CLASSES"
607 nigel 63 .rs
608     .sp
609     An opening square bracket introduces a character class, terminated by a closing
610     square bracket. A closing square bracket on its own is not special. If a
611     closing square bracket is required as a member of the class, it should be the
612     first data character in the class (after an initial circumflex, if present) or
613     escaped with a backslash.
614 nigel 75 .P
615 nigel 63 A character class matches a single character in the subject. In UTF-8 mode, the
616     character may occupy more than one byte. A matched character must be in the set
617     of characters defined by the class, unless the first character in the class
618     definition is a circumflex, in which case the subject character must not be in
619     the set defined by the class. If a circumflex is actually required as a member
620     of the class, ensure it is not the first character, or escape it with a
621     backslash.
622 nigel 75 .P
623 nigel 63 For example, the character class [aeiou] matches any lower case vowel, while
624     [^aeiou] matches any character that is not a lower case vowel. Note that a
625 nigel 75 circumflex is just a convenient notation for specifying the characters that
626     are in the class by enumerating those that are not. A class that starts with a
627     circumflex is not an assertion: it still consumes a character from the subject
628     string, and therefore it fails if the current pointer is at the end of the
629     string.
630     .P
631 nigel 63 In UTF-8 mode, characters with values greater than 255 can be included in a
632 nigel 75 class as a literal string of bytes, or by using the \ex{ escaping mechanism.
633     .P
634 nigel 63 When caseless matching is set, any letters in a class represent both their
635     upper case and lower case versions, so for example, a caseless [aeiou] matches
636     "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
637 nigel 77 caseful version would. In UTF-8 mode, PCRE always understands the concept of
638     case for characters whose values are less than 128, so caseless matching is
639     always possible. For characters with higher values, the concept of case is
640     supported if PCRE is compiled with Unicode property support, but not otherwise.
641     If you want to use caseless matching for characters 128 and above, you must
642     ensure that PCRE is compiled with Unicode property support as well as with
643     UTF-8 support.
644 nigel 75 .P
645 nigel 63 The newline character is never treated in any special way in character classes,
646     whatever the setting of the PCRE_DOTALL or PCRE_MULTILINE options is. A class
647     such as [^a] will always match a newline.
648 nigel 75 .P
649 nigel 63 The minus (hyphen) character can be used to specify a range of characters in a
650     character class. For example, [d-m] matches any letter between d and m,
651     inclusive. If a minus character is required in a class, it must be escaped with
652     a backslash or appear in a position where it cannot be interpreted as
653     indicating a range, typically as the first or last character in the class.
654 nigel 75 .P
655 nigel 63 It is not possible to have the literal character "]" as the end character of a
656     range. A pattern such as [W-]46] is interpreted as a class of two characters
657     ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
658     "-46]". However, if the "]" is escaped with a backslash it is interpreted as
659 nigel 75 the end of range, so [W-\e]46] is interpreted as a class containing a range
660     followed by two other characters. The octal or hexadecimal representation of
661     "]" can also be used to end a range.
662     .P
663 nigel 63 Ranges operate in the collating sequence of character values. They can also be
664 nigel 75 used for characters specified numerically, for example [\e000-\e037]. In UTF-8
665 nigel 63 mode, ranges can include characters whose values are greater than 255, for
666 nigel 75 example [\ex{100}-\ex{2ff}].
667     .P
668 nigel 63 If a range that includes letters is used when caseless matching is set, it
669     matches the letters in either case. For example, [W-c] is equivalent to
670 nigel 75 [][\e\e^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
671     tables for the "fr_FR" locale are in use, [\exc8-\excb] matches accented E
672     characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
673     characters with values greater than 128 only when it is compiled with Unicode
674     property support.
675     .P
676     The character types \ed, \eD, \ep, \eP, \es, \eS, \ew, and \eW may also appear
677     in a character class, and add the characters that they match to the class. For
678     example, [\edABCDEF] matches any hexadecimal digit. A circumflex can
679 nigel 63 conveniently be used with the upper case character types to specify a more
680     restricted set of characters than the matching lower case type. For example,
681 nigel 75 the class [^\eW_] matches any letter or digit, but not underscore.
682     .P
683     The only metacharacters that are recognized in character classes are backslash,
684     hyphen (only where it can be interpreted as specifying a range), circumflex
685     (only at the start), opening square bracket (only when it can be interpreted as
686     introducing a POSIX class name - see the next section), and the terminating
687     closing square bracket. However, escaping other non-alphanumeric characters
688     does no harm.
689     .
690     .
691     .SH "POSIX CHARACTER CLASSES"
692 nigel 63 .rs
693     .sp
694 nigel 75 Perl supports the POSIX notation for character classes. This uses names
695 nigel 63 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
696     this notation. For example,
697 nigel 75 .sp
698 nigel 63 [01[:alpha:]%]
699 nigel 75 .sp
700 nigel 63 matches "0", "1", any alphabetic character, or "%". The supported class names
701     are
702 nigel 75 .sp
703 nigel 63 alnum letters and digits
704     alpha letters
705     ascii character codes 0 - 127
706     blank space or tab only
707     cntrl control characters
708 nigel 75 digit decimal digits (same as \ed)
709 nigel 63 graph printing characters, excluding space
710     lower lower case letters
711     print printing characters, including space
712     punct printing characters, excluding letters and digits
713 nigel 75 space white space (not quite the same as \es)
714 nigel 63 upper upper case letters
715 nigel 75 word "word" characters (same as \ew)
716 nigel 63 xdigit hexadecimal digits
717 nigel 75 .sp
718 nigel 63 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
719     space (32). Notice that this list includes the VT character (code 11). This
720 nigel 75 makes "space" different to \es, which does not include VT (for Perl
721 nigel 63 compatibility).
722 nigel 75 .P
723 nigel 63 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
724     5.8. Another Perl extension is negation, which is indicated by a ^ character
725     after the colon. For example,
726 nigel 75 .sp
727 nigel 63 [12[:^digit:]]
728 nigel 75 .sp
729 nigel 63 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
730     syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
731     supported, and an error is given if they are encountered.
732 nigel 75 .P
733     In UTF-8 mode, characters with values greater than 128 do not match any of
734 nigel 63 the POSIX character classes.
735 nigel 75 .
736     .
737     .SH "VERTICAL BAR"
738 nigel 63 .rs
739     .sp
740     Vertical bar characters are used to separate alternative patterns. For example,
741     the pattern
742 nigel 75 .sp
743 nigel 63 gilbert|sullivan
744 nigel 75 .sp
745 nigel 63 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
746     and an empty alternative is permitted (matching the empty string).
747     The matching process tries each alternative in turn, from left to right,
748     and the first one that succeeds is used. If the alternatives are within a
749 nigel 75 subpattern
750     .\" HTML <a href="#subpattern">
751     .\" </a>
752     (defined below),
753     .\"
754     "succeeds" means matching the rest of the main pattern as well as the
755     alternative in the subpattern.
756     .
757     .
758     .SH "INTERNAL OPTION SETTING"
759 nigel 63 .rs
760     .sp
761     The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
762     PCRE_EXTENDED options can be changed from within the pattern by a sequence of
763     Perl option letters enclosed between "(?" and ")". The option letters are
764 nigel 75 .sp
765 nigel 63 i for PCRE_CASELESS
766     m for PCRE_MULTILINE
767     s for PCRE_DOTALL
768     x for PCRE_EXTENDED
769 nigel 75 .sp
770 nigel 63 For example, (?im) sets caseless, multiline matching. It is also possible to
771     unset these options by preceding the letter with a hyphen, and a combined
772     setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
773     PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
774     permitted. If a letter appears both before and after the hyphen, the option is
775     unset.
776 nigel 75 .P
777 nigel 63 When an option change occurs at top level (that is, not inside subpattern
778     parentheses), the change applies to the remainder of the pattern that follows.
779     If the change is placed right at the start of a pattern, PCRE extracts it into
780     the global options (and it will therefore show up in data extracted by the
781 nigel 75 \fBpcre_fullinfo()\fP function).
782     .P
783 nigel 63 An option change within a subpattern affects only that part of the current
784     pattern that follows it, so
785 nigel 75 .sp
786 nigel 63 (a(?i)b)c
787 nigel 75 .sp
788 nigel 63 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
789     By this means, options can be made to have different settings in different
790     parts of the pattern. Any changes made in one alternative do carry on
791     into subsequent branches within the same subpattern. For example,
792 nigel 75 .sp
793 nigel 63 (a(?i)b|c)
794 nigel 75 .sp
795 nigel 63 matches "ab", "aB", "c", and "C", even though when matching "C" the first
796     branch is abandoned before the option setting. This is because the effects of
797     option settings happen at compile time. There would be some very weird
798     behaviour otherwise.
799 nigel 75 .P
800 nigel 63 The PCRE-specific options PCRE_UNGREEDY and PCRE_EXTRA can be changed in the
801     same way as the Perl-compatible options by using the characters U and X
802     respectively. The (?X) flag setting is special in that it must always occur
803     earlier in the pattern than any of the additional features it turns on, even
804 nigel 75 when it is at top level. It is best to put it at the start.
805     .
806     .
807     .\" HTML <a name="subpattern"></a>
808 nigel 63 .SH SUBPATTERNS
809     .rs
810     .sp
811     Subpatterns are delimited by parentheses (round brackets), which can be nested.
812 nigel 75 Turning part of a pattern into a subpattern does two things:
813     .sp
814 nigel 63 1. It localizes a set of alternatives. For example, the pattern
815 nigel 75 .sp
816 nigel 63 cat(aract|erpillar|)
817 nigel 75 .sp
818 nigel 63 matches one of the words "cat", "cataract", or "caterpillar". Without the
819     parentheses, it would match "cataract", "erpillar" or the empty string.
820 nigel 75 .sp
821     2. It sets up the subpattern as a capturing subpattern. This means that, when
822     the whole pattern matches, that portion of the subject string that matched the
823     subpattern is passed back to the caller via the \fIovector\fP argument of
824     \fBpcre_exec()\fP. Opening parentheses are counted from left to right (starting
825     from 1) to obtain numbers for the capturing subpatterns.
826     .P
827 nigel 63 For example, if the string "the red king" is matched against the pattern
828 nigel 75 .sp
829 nigel 63 the ((red|white) (king|queen))
830 nigel 75 .sp
831 nigel 63 the captured substrings are "red king", "red", and "king", and are numbered 1,
832     2, and 3, respectively.
833 nigel 75 .P
834 nigel 63 The fact that plain parentheses fulfil two functions is not always helpful.
835     There are often times when a grouping subpattern is required without a
836     capturing requirement. If an opening parenthesis is followed by a question mark
837     and a colon, the subpattern does not do any capturing, and is not counted when
838     computing the number of any subsequent capturing subpatterns. For example, if
839     the string "the white queen" is matched against the pattern
840 nigel 75 .sp
841 nigel 63 the ((?:red|white) (king|queen))
842 nigel 75 .sp
843 nigel 63 the captured substrings are "white queen" and "queen", and are numbered 1 and
844     2. The maximum number of capturing subpatterns is 65535, and the maximum depth
845     of nesting of all subpatterns, both capturing and non-capturing, is 200.
846 nigel 75 .P
847 nigel 63 As a convenient shorthand, if any option settings are required at the start of
848     a non-capturing subpattern, the option letters may appear between the "?" and
849     the ":". Thus the two patterns
850 nigel 75 .sp
851 nigel 63 (?i:saturday|sunday)
852     (?:(?i)saturday|sunday)
853 nigel 75 .sp
854 nigel 63 match exactly the same set of strings. Because alternative branches are tried
855     from left to right, and options are not reset until the end of the subpattern
856     is reached, an option setting in one branch does affect subsequent branches, so
857     the above patterns match "SUNDAY" as well as "Saturday".
858 nigel 75 .
859     .
860     .SH "NAMED SUBPATTERNS"
861 nigel 63 .rs
862     .sp
863     Identifying capturing parentheses by number is simple, but it can be very hard
864     to keep track of the numbers in complicated regular expressions. Furthermore,
865 nigel 75 if an expression is modified, the numbers may change. To help with this
866 nigel 63 difficulty, PCRE supports the naming of subpatterns, something that Perl does
867     not provide. The Python syntax (?P<name>...) is used. Names consist of
868     alphanumeric characters and underscores, and must be unique within a pattern.
869 nigel 75 .P
870 nigel 63 Named capturing parentheses are still allocated numbers as well as names. The
871     PCRE API provides function calls for extracting the name-to-number translation
872 nigel 75 table from a compiled pattern. There is also a convenience function for
873     extracting a captured substring by name. For further details see the
874 nigel 63 .\" HREF
875 nigel 75 \fBpcreapi\fP
876 nigel 63 .\"
877     documentation.
878 nigel 75 .
879     .
880 nigel 63 .SH REPETITION
881     .rs
882     .sp
883     Repetition is specified by quantifiers, which can follow any of the following
884     items:
885 nigel 75 .sp
886 nigel 63 a literal data character
887     the . metacharacter
888 nigel 75 the \eC escape sequence
889     the \eX escape sequence (in UTF-8 mode with Unicode properties)
890     an escape such as \ed that matches a single character
891 nigel 63 a character class
892     a back reference (see next section)
893     a parenthesized subpattern (unless it is an assertion)
894 nigel 75 .sp
895 nigel 63 The general repetition quantifier specifies a minimum and maximum number of
896     permitted matches, by giving the two numbers in curly brackets (braces),
897     separated by a comma. The numbers must be less than 65536, and the first must
898     be less than or equal to the second. For example:
899 nigel 75 .sp
900 nigel 63 z{2,4}
901 nigel 75 .sp
902 nigel 63 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
903     character. If the second number is omitted, but the comma is present, there is
904     no upper limit; if the second number and the comma are both omitted, the
905     quantifier specifies an exact number of required matches. Thus
906 nigel 75 .sp
907 nigel 63 [aeiou]{3,}
908 nigel 75 .sp
909 nigel 63 matches at least 3 successive vowels, but may match many more, while
910 nigel 75 .sp
911     \ed{8}
912     .sp
913 nigel 63 matches exactly 8 digits. An opening curly bracket that appears in a position
914     where a quantifier is not allowed, or one that does not match the syntax of a
915     quantifier, is taken as a literal character. For example, {,6} is not a
916     quantifier, but a literal string of four characters.
917 nigel 75 .P
918 nigel 63 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
919 nigel 75 bytes. Thus, for example, \ex{100}{2} matches two UTF-8 characters, each of
920     which is represented by a two-byte sequence. Similarly, when Unicode property
921     support is available, \eX{3} matches three Unicode extended sequences, each of
922     which may be several bytes long (and they may be of different lengths).
923     .P
924 nigel 63 The quantifier {0} is permitted, causing the expression to behave as if the
925     previous item and the quantifier were not present.
926 nigel 75 .P
927 nigel 63 For convenience (and historical compatibility) the three most common
928     quantifiers have single-character abbreviations:
929 nigel 75 .sp
930 nigel 63 * is equivalent to {0,}
931     + is equivalent to {1,}
932     ? is equivalent to {0,1}
933 nigel 75 .sp
934 nigel 63 It is possible to construct infinite loops by following a subpattern that can
935     match no characters with a quantifier that has no upper limit, for example:
936 nigel 75 .sp
937 nigel 63 (a?)*
938 nigel 75 .sp
939 nigel 63 Earlier versions of Perl and PCRE used to give an error at compile time for
940     such patterns. However, because there are cases where this can be useful, such
941     patterns are now accepted, but if any repetition of the subpattern does in fact
942     match no characters, the loop is forcibly broken.
943 nigel 75 .P
944 nigel 63 By default, the quantifiers are "greedy", that is, they match as much as
945     possible (up to the maximum number of permitted times), without causing the
946     rest of the pattern to fail. The classic example of where this gives problems
947 nigel 75 is in trying to match comments in C programs. These appear between /* and */
948     and within the comment, individual * and / characters may appear. An attempt to
949     match C comments by applying the pattern
950     .sp
951     /\e*.*\e*/
952     .sp
953 nigel 63 to the string
954 nigel 75 .sp
955     /* first comment */ not comment /* second comment */
956     .sp
957 nigel 63 fails, because it matches the entire string owing to the greediness of the .*
958     item.
959 nigel 75 .P
960 nigel 63 However, if a quantifier is followed by a question mark, it ceases to be
961     greedy, and instead matches the minimum number of times possible, so the
962     pattern
963 nigel 75 .sp
964     /\e*.*?\e*/
965     .sp
966 nigel 63 does the right thing with the C comments. The meaning of the various
967     quantifiers is not otherwise changed, just the preferred number of matches.
968     Do not confuse this use of question mark with its use as a quantifier in its
969     own right. Because it has two uses, it can sometimes appear doubled, as in
970 nigel 75 .sp
971     \ed??\ed
972     .sp
973 nigel 63 which matches one digit by preference, but can match two if that is the only
974     way the rest of the pattern matches.
975 nigel 75 .P
976 nigel 63 If the PCRE_UNGREEDY option is set (an option which is not available in Perl),
977     the quantifiers are not greedy by default, but individual ones can be made
978     greedy by following them with a question mark. In other words, it inverts the
979     default behaviour.
980 nigel 75 .P
981 nigel 63 When a parenthesized subpattern is quantified with a minimum repeat count that
982 nigel 75 is greater than 1 or with a limited maximum, more memory is required for the
983 nigel 63 compiled pattern, in proportion to the size of the minimum or maximum.
984 nigel 75 .P
985 nigel 63 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
986     to Perl's /s) is set, thus allowing the . to match newlines, the pattern is
987     implicitly anchored, because whatever follows will be tried against every
988     character position in the subject string, so there is no point in retrying the
989     overall match at any position after the first. PCRE normally treats such a
990 nigel 75 pattern as though it were preceded by \eA.
991     .P
992 nigel 63 In cases where it is known that the subject string contains no newlines, it is
993     worth setting PCRE_DOTALL in order to obtain this optimization, or
994     alternatively using ^ to indicate anchoring explicitly.
995 nigel 75 .P
996 nigel 63 However, there is one situation where the optimization cannot be used. When .*
997     is inside capturing parentheses that are the subject of a backreference
998     elsewhere in the pattern, a match at the start may fail, and a later one
999     succeed. Consider, for example:
1000 nigel 75 .sp
1001     (.*)abc\e1
1002     .sp
1003 nigel 63 If the subject is "xyz123abc123" the match point is the fourth character. For
1004     this reason, such a pattern is not implicitly anchored.
1005 nigel 75 .P
1006 nigel 63 When a capturing subpattern is repeated, the value captured is the substring
1007     that matched the final iteration. For example, after
1008 nigel 75 .sp
1009     (tweedle[dume]{3}\es*)+
1010     .sp
1011 nigel 63 has matched "tweedledum tweedledee" the value of the captured substring is
1012     "tweedledee". However, if there are nested capturing subpatterns, the
1013     corresponding captured values may have been set in previous iterations. For
1014     example, after
1015 nigel 75 .sp
1016 nigel 63 /(a|(b))+/
1017 nigel 75 .sp
1018 nigel 63 matches "aba" the value of the second captured substring is "b".
1019 nigel 75 .
1020     .
1021     .\" HTML <a name="atomicgroup"></a>
1022     .SH "ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS"
1023 nigel 63 .rs
1024     .sp
1025     With both maximizing and minimizing repetition, failure of what follows
1026     normally causes the repeated item to be re-evaluated to see if a different
1027     number of repeats allows the rest of the pattern to match. Sometimes it is
1028     useful to prevent this, either to change the nature of the match, or to cause
1029     it fail earlier than it otherwise might, when the author of the pattern knows
1030     there is no point in carrying on.
1031 nigel 75 .P
1032     Consider, for example, the pattern \ed+foo when applied to the subject line
1033     .sp
1034 nigel 63 123456bar
1035 nigel 75 .sp
1036 nigel 63 After matching all 6 digits and then failing to match "foo", the normal
1037 nigel 75 action of the matcher is to try again with only 5 digits matching the \ed+
1038 nigel 63 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1039     (a term taken from Jeffrey Friedl's book) provides the means for specifying
1040     that once a subpattern has matched, it is not to be re-evaluated in this way.
1041 nigel 75 .P
1042 nigel 63 If we use atomic grouping for the previous example, the matcher would give up
1043     immediately on failing to match "foo" the first time. The notation is a kind of
1044     special parenthesis, starting with (?> as in this example:
1045 nigel 75 .sp
1046     (?>\ed+)foo
1047     .sp
1048 nigel 63 This kind of parenthesis "locks up" the part of the pattern it contains once
1049     it has matched, and a failure further into the pattern is prevented from
1050     backtracking into it. Backtracking past it to previous items, however, works as
1051     normal.
1052 nigel 75 .P
1053 nigel 63 An alternative description is that a subpattern of this type matches the string
1054     of characters that an identical standalone pattern would match, if anchored at
1055     the current point in the subject string.
1056 nigel 75 .P
1057 nigel 63 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1058     the above example can be thought of as a maximizing repeat that must swallow
1059 nigel 75 everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1060 nigel 63 number of digits they match in order to make the rest of the pattern match,
1061 nigel 75 (?>\ed+) can only match an entire sequence of digits.
1062     .P
1063 nigel 63 Atomic groups in general can of course contain arbitrarily complicated
1064     subpatterns, and can be nested. However, when the subpattern for an atomic
1065     group is just a single repeated item, as in the example above, a simpler
1066     notation, called a "possessive quantifier" can be used. This consists of an
1067     additional + character following a quantifier. Using this notation, the
1068     previous example can be rewritten as
1069 nigel 75 .sp
1070     \ed++foo
1071     .sp
1072 nigel 63 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1073     option is ignored. They are a convenient notation for the simpler forms of
1074     atomic group. However, there is no difference in the meaning or processing of a
1075     possessive quantifier and the equivalent atomic group.
1076 nigel 75 .P
1077 nigel 63 The possessive quantifier syntax is an extension to the Perl syntax. It
1078     originates in Sun's Java package.
1079 nigel 75 .P
1080 nigel 63 When a pattern contains an unlimited repeat inside a subpattern that can itself
1081     be repeated an unlimited number of times, the use of an atomic group is the
1082     only way to avoid some failing matches taking a very long time indeed. The
1083     pattern
1084 nigel 75 .sp
1085     (\eD+|<\ed+>)*[!?]
1086     .sp
1087 nigel 63 matches an unlimited number of substrings that either consist of non-digits, or
1088     digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1089     quickly. However, if it is applied to
1090 nigel 75 .sp
1091 nigel 63 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1092 nigel 75 .sp
1093 nigel 63 it takes a long time before reporting failure. This is because the string can
1094 nigel 75 be divided between the internal \eD+ repeat and the external * repeat in a
1095     large number of ways, and all have to be tried. (The example uses [!?] rather
1096     than a single character at the end, because both PCRE and Perl have an
1097     optimization that allows for fast failure when a single character is used. They
1098     remember the last single character that is required for a match, and fail early
1099     if it is not present in the string.) If the pattern is changed so that it uses
1100     an atomic group, like this:
1101     .sp
1102     ((?>\eD+)|<\ed+>)*[!?]
1103     .sp
1104 nigel 63 sequences of non-digits cannot be broken, and failure happens quickly.
1105 nigel 75 .
1106     .
1107     .\" HTML <a name="backreferences"></a>
1108     .SH "BACK REFERENCES"
1109 nigel 63 .rs
1110     .sp
1111     Outside a character class, a backslash followed by a digit greater than 0 (and
1112     possibly further digits) is a back reference to a capturing subpattern earlier
1113     (that is, to its left) in the pattern, provided there have been that many
1114     previous capturing left parentheses.
1115 nigel 75 .P
1116 nigel 63 However, if the decimal number following the backslash is less than 10, it is
1117     always taken as a back reference, and causes an error only if there are not
1118     that many capturing left parentheses in the entire pattern. In other words, the
1119     parentheses that are referenced need not be to the left of the reference for
1120 nigel 75 numbers less than 10. See the subsection entitled "Non-printing characters"
1121     .\" HTML <a href="#digitsafterbackslash">
1122     .\" </a>
1123     above
1124     .\"
1125     for further details of the handling of digits following a backslash.
1126     .P
1127 nigel 63 A back reference matches whatever actually matched the capturing subpattern in
1128     the current subject string, rather than anything matching the subpattern
1129     itself (see
1130     .\" HTML <a href="#subpatternsassubroutines">
1131     .\" </a>
1132     "Subpatterns as subroutines"
1133     .\"
1134     below for a way of doing that). So the pattern
1135 nigel 75 .sp
1136     (sens|respons)e and \e1ibility
1137     .sp
1138 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1139     "sense and responsibility". If caseful matching is in force at the time of the
1140     back reference, the case of letters is relevant. For example,
1141 nigel 75 .sp
1142     ((?i)rah)\es+\e1
1143     .sp
1144 nigel 63 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1145     capturing subpattern is matched caselessly.
1146 nigel 75 .P
1147 nigel 63 Back references to named subpatterns use the Python syntax (?P=name). We could
1148     rewrite the above example as follows:
1149 nigel 75 .sp
1150     (?<p1>(?i)rah)\es+(?P=p1)
1151     .sp
1152 nigel 63 There may be more than one back reference to the same subpattern. If a
1153     subpattern has not actually been used in a particular match, any back
1154     references to it always fail. For example, the pattern
1155 nigel 75 .sp
1156     (a|(bc))\e2
1157     .sp
1158 nigel 63 always fails if it starts to match "a" rather than "bc". Because there may be
1159     many capturing parentheses in a pattern, all digits following the backslash are
1160     taken as part of a potential back reference number. If the pattern continues
1161     with a digit character, some delimiter must be used to terminate the back
1162     reference. If the PCRE_EXTENDED option is set, this can be whitespace.
1163 nigel 75 Otherwise an empty comment (see
1164     .\" HTML <a href="#comments">
1165     .\" </a>
1166     "Comments"
1167     .\"
1168     below) can be used.
1169     .P
1170 nigel 63 A back reference that occurs inside the parentheses to which it refers fails
1171 nigel 75 when the subpattern is first used, so, for example, (a\e1) never matches.
1172 nigel 63 However, such references can be useful inside repeated subpatterns. For
1173     example, the pattern
1174 nigel 75 .sp
1175     (a|b\e1)+
1176     .sp
1177 nigel 63 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1178     the subpattern, the back reference matches the character string corresponding
1179     to the previous iteration. In order for this to work, the pattern must be such
1180     that the first iteration does not need to match the back reference. This can be
1181     done using alternation, as in the example above, or by a quantifier with a
1182     minimum of zero.
1183 nigel 75 .
1184     .
1185     .\" HTML <a name="bigassertions"></a>
1186 nigel 63 .SH ASSERTIONS
1187     .rs
1188     .sp
1189     An assertion is a test on the characters following or preceding the current
1190     matching point that does not actually consume any characters. The simple
1191 nigel 75 assertions coded as \eb, \eB, \eA, \eG, \eZ, \ez, ^ and $ are described
1192     .\" HTML <a href="#smallassertions">
1193     .\" </a>
1194     above.
1195     .\"
1196     .P
1197 nigel 63 More complicated assertions are coded as subpatterns. There are two kinds:
1198     those that look ahead of the current position in the subject string, and those
1199 nigel 75 that look behind it. An assertion subpattern is matched in the normal way,
1200     except that it does not cause the current matching position to be changed.
1201     .P
1202     Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1203     because it makes no sense to assert the same thing several times. If any kind
1204     of assertion contains capturing subpatterns within it, these are counted for
1205     the purposes of numbering the capturing subpatterns in the whole pattern.
1206     However, substring capturing is carried out only for positive assertions,
1207     because it does not make sense for negative assertions.
1208     .
1209     .
1210     .SS "Lookahead assertions"
1211     .rs
1212     .sp
1213     Lookahead assertions start
1214 nigel 63 with (?= for positive assertions and (?! for negative assertions. For example,
1215 nigel 75 .sp
1216     \ew+(?=;)
1217     .sp
1218 nigel 63 matches a word followed by a semicolon, but does not include the semicolon in
1219     the match, and
1220 nigel 75 .sp
1221 nigel 63 foo(?!bar)
1222 nigel 75 .sp
1223 nigel 63 matches any occurrence of "foo" that is not followed by "bar". Note that the
1224     apparently similar pattern
1225 nigel 75 .sp
1226 nigel 63 (?!foo)bar
1227 nigel 75 .sp
1228 nigel 63 does not find an occurrence of "bar" that is preceded by something other than
1229     "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1230     (?!foo) is always true when the next three characters are "bar". A
1231 nigel 75 lookbehind assertion is needed to achieve the other effect.
1232     .P
1233 nigel 63 If you want to force a matching failure at some point in a pattern, the most
1234     convenient way to do it is with (?!) because an empty string always matches, so
1235     an assertion that requires there not to be an empty string must always fail.
1236 nigel 75 .
1237     .
1238     .\" HTML <a name="lookbehind"></a>
1239     .SS "Lookbehind assertions"
1240     .rs
1241     .sp
1242 nigel 63 Lookbehind assertions start with (?<= for positive assertions and (?<! for
1243     negative assertions. For example,
1244 nigel 75 .sp
1245 nigel 63 (?<!foo)bar
1246 nigel 75 .sp
1247 nigel 63 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1248     a lookbehind assertion are restricted such that all the strings it matches must
1249     have a fixed length. However, if there are several alternatives, they do not
1250     all have to have the same fixed length. Thus
1251 nigel 75 .sp
1252 nigel 63 (?<=bullock|donkey)
1253 nigel 75 .sp
1254 nigel 63 is permitted, but
1255 nigel 75 .sp
1256 nigel 63 (?<!dogs?|cats?)
1257 nigel 75 .sp
1258 nigel 63 causes an error at compile time. Branches that match different length strings
1259     are permitted only at the top level of a lookbehind assertion. This is an
1260     extension compared with Perl (at least for 5.8), which requires all branches to
1261     match the same length of string. An assertion such as
1262 nigel 75 .sp
1263 nigel 63 (?<=ab(c|de))
1264 nigel 75 .sp
1265 nigel 63 is not permitted, because its single top-level branch can match two different
1266     lengths, but it is acceptable if rewritten to use two top-level branches:
1267 nigel 75 .sp
1268 nigel 63 (?<=abc|abde)
1269 nigel 75 .sp
1270 nigel 63 The implementation of lookbehind assertions is, for each alternative, to
1271     temporarily move the current position back by the fixed width and then try to
1272     match. If there are insufficient characters before the current position, the
1273     match is deemed to fail.
1274 nigel 75 .P
1275     PCRE does not allow the \eC escape (which matches a single byte in UTF-8 mode)
1276 nigel 63 to appear in lookbehind assertions, because it makes it impossible to calculate
1277 nigel 75 the length of the lookbehind. The \eX escape, which can match different numbers
1278     of bytes, is also not permitted.
1279     .P
1280 nigel 63 Atomic groups can be used in conjunction with lookbehind assertions to specify
1281     efficient matching at the end of the subject string. Consider a simple pattern
1282     such as
1283 nigel 75 .sp
1284 nigel 63 abcd$
1285 nigel 75 .sp
1286 nigel 63 when applied to a long string that does not match. Because matching proceeds
1287     from left to right, PCRE will look for each "a" in the subject and then see if
1288     what follows matches the rest of the pattern. If the pattern is specified as
1289 nigel 75 .sp
1290 nigel 63 ^.*abcd$
1291 nigel 75 .sp
1292 nigel 63 the initial .* matches the entire string at first, but when this fails (because
1293     there is no following "a"), it backtracks to match all but the last character,
1294     then all but the last two characters, and so on. Once again the search for "a"
1295     covers the entire string, from right to left, so we are no better off. However,
1296     if the pattern is written as
1297 nigel 75 .sp
1298 nigel 63 ^(?>.*)(?<=abcd)
1299 nigel 75 .sp
1300     or, equivalently, using the possessive quantifier syntax,
1301     .sp
1302 nigel 63 ^.*+(?<=abcd)
1303 nigel 75 .sp
1304 nigel 63 there can be no backtracking for the .* item; it can match only the entire
1305     string. The subsequent lookbehind assertion does a single test on the last four
1306     characters. If it fails, the match fails immediately. For long strings, this
1307     approach makes a significant difference to the processing time.
1308 nigel 75 .
1309     .
1310     .SS "Using multiple assertions"
1311     .rs
1312     .sp
1313 nigel 63 Several assertions (of any sort) may occur in succession. For example,
1314 nigel 75 .sp
1315     (?<=\ed{3})(?<!999)foo
1316     .sp
1317 nigel 63 matches "foo" preceded by three digits that are not "999". Notice that each of
1318     the assertions is applied independently at the same point in the subject
1319     string. First there is a check that the previous three characters are all
1320     digits, and then there is a check that the same three characters are not "999".
1321 nigel 75 This pattern does \fInot\fP match "foo" preceded by six characters, the first
1322 nigel 63 of which are digits and the last three of which are not "999". For example, it
1323     doesn't match "123abcfoo". A pattern to do that is
1324 nigel 75 .sp
1325     (?<=\ed{3}...)(?<!999)foo
1326     .sp
1327 nigel 63 This time the first assertion looks at the preceding six characters, checking
1328     that the first three are digits, and then the second assertion checks that the
1329     preceding three characters are not "999".
1330 nigel 75 .P
1331 nigel 63 Assertions can be nested in any combination. For example,
1332 nigel 75 .sp
1333 nigel 63 (?<=(?<!foo)bar)baz
1334 nigel 75 .sp
1335 nigel 63 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1336     preceded by "foo", while
1337 nigel 75 .sp
1338     (?<=\ed{3}(?!999)...)foo
1339     .sp
1340     is another pattern that matches "foo" preceded by three digits and any three
1341 nigel 63 characters that are not "999".
1342 nigel 75 .
1343     .
1344     .SH "CONDITIONAL SUBPATTERNS"
1345 nigel 63 .rs
1346     .sp
1347     It is possible to cause the matching process to obey a subpattern
1348     conditionally or to choose between two alternative subpatterns, depending on
1349     the result of an assertion, or whether a previous capturing subpattern matched
1350     or not. The two possible forms of conditional subpattern are
1351 nigel 75 .sp
1352 nigel 63 (?(condition)yes-pattern)
1353     (?(condition)yes-pattern|no-pattern)
1354 nigel 75 .sp
1355 nigel 63 If the condition is satisfied, the yes-pattern is used; otherwise the
1356     no-pattern (if present) is used. If there are more than two alternatives in the
1357     subpattern, a compile-time error occurs.
1358 nigel 75 .P
1359 nigel 63 There are three kinds of condition. If the text between the parentheses
1360     consists of a sequence of digits, the condition is satisfied if the capturing
1361     subpattern of that number has previously matched. The number must be greater
1362     than zero. Consider the following pattern, which contains non-significant white
1363     space to make it more readable (assume the PCRE_EXTENDED option) and to divide
1364     it into three parts for ease of discussion:
1365 nigel 75 .sp
1366     ( \e( )? [^()]+ (?(1) \e) )
1367     .sp
1368 nigel 63 The first part matches an optional opening parenthesis, and if that
1369     character is present, sets it as the first captured substring. The second part
1370     matches one or more characters that are not parentheses. The third part is a
1371     conditional subpattern that tests whether the first set of parentheses matched
1372     or not. If they did, that is, if subject started with an opening parenthesis,
1373     the condition is true, and so the yes-pattern is executed and a closing
1374     parenthesis is required. Otherwise, since no-pattern is not present, the
1375     subpattern matches nothing. In other words, this pattern matches a sequence of
1376     non-parentheses, optionally enclosed in parentheses.
1377 nigel 75 .P
1378 nigel 63 If the condition is the string (R), it is satisfied if a recursive call to the
1379     pattern or subpattern has been made. At "top level", the condition is false.
1380     This is a PCRE extension. Recursive patterns are described in the next section.
1381 nigel 75 .P
1382 nigel 63 If the condition is not a sequence of digits or (R), it must be an assertion.
1383     This may be a positive or negative lookahead or lookbehind assertion. Consider
1384     this pattern, again containing non-significant white space, and with the two
1385     alternatives on the second line:
1386 nigel 75 .sp
1387 nigel 63 (?(?=[^a-z]*[a-z])
1388 nigel 75 \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
1389     .sp
1390 nigel 63 The condition is a positive lookahead assertion that matches an optional
1391     sequence of non-letters followed by a letter. In other words, it tests for the
1392     presence of at least one letter in the subject. If a letter is found, the
1393     subject is matched against the first alternative; otherwise it is matched
1394     against the second. This pattern matches strings in one of the two forms
1395     dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
1396 nigel 75 .
1397     .
1398     .\" HTML <a name="comments"></a>
1399 nigel 63 .SH COMMENTS
1400     .rs
1401     .sp
1402 nigel 75 The sequence (?# marks the start of a comment that continues up to the next
1403 nigel 63 closing parenthesis. Nested parentheses are not permitted. The characters
1404     that make up a comment play no part in the pattern matching at all.
1405 nigel 75 .P
1406 nigel 63 If the PCRE_EXTENDED option is set, an unescaped # character outside a
1407     character class introduces a comment that continues up to the next newline
1408     character in the pattern.
1409 nigel 75 .
1410     .
1411     .SH "RECURSIVE PATTERNS"
1412 nigel 63 .rs
1413     .sp
1414     Consider the problem of matching a string in parentheses, allowing for
1415     unlimited nested parentheses. Without the use of recursion, the best that can
1416     be done is to use a pattern that matches up to some fixed depth of nesting. It
1417 nigel 75 is not possible to handle an arbitrary nesting depth. Perl provides a facility
1418     that allows regular expressions to recurse (amongst other things). It does this
1419     by interpolating Perl code in the expression at run time, and the code can
1420     refer to the expression itself. A Perl pattern to solve the parentheses problem
1421     can be created like this:
1422     .sp
1423     $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
1424     .sp
1425 nigel 63 The (?p{...}) item interpolates Perl code at run time, and in this case refers
1426     recursively to the pattern in which it appears. Obviously, PCRE cannot support
1427     the interpolation of Perl code. Instead, it supports some special syntax for
1428     recursion of the entire pattern, and also for individual subpattern recursion.
1429 nigel 75 .P
1430 nigel 63 The special item that consists of (? followed by a number greater than zero and
1431     a closing parenthesis is a recursive call of the subpattern of the given
1432     number, provided that it occurs inside that subpattern. (If not, it is a
1433     "subroutine" call, which is described in the next section.) The special item
1434     (?R) is a recursive call of the entire regular expression.
1435 nigel 75 .P
1436 nigel 87 A recursive subpattern call is always treated as an atomic group. That is, once
1437     it has matched some of the subject string, it is never re-entered, even if
1438     it contains untried alternatives and there is a subsequent matching failure.
1439     .P
1440     This PCRE pattern solves the nested parentheses problem (assume the
1441     PCRE_EXTENDED option is set so that white space is ignored):
1442 nigel 75 .sp
1443     \e( ( (?>[^()]+) | (?R) )* \e)
1444     .sp
1445 nigel 63 First it matches an opening parenthesis. Then it matches any number of
1446     substrings which can either be a sequence of non-parentheses, or a recursive
1447 nigel 87 match of the pattern itself (that is, a correctly parenthesized substring).
1448 nigel 63 Finally there is a closing parenthesis.
1449 nigel 75 .P
1450 nigel 63 If this were part of a larger pattern, you would not want to recurse the entire
1451     pattern, so instead you could use this:
1452 nigel 75 .sp
1453     ( \e( ( (?>[^()]+) | (?1) )* \e) )
1454     .sp
1455 nigel 63 We have put the pattern into parentheses, and caused the recursion to refer to
1456     them instead of the whole pattern. In a larger pattern, keeping track of
1457     parenthesis numbers can be tricky. It may be more convenient to use named
1458     parentheses instead. For this, PCRE uses (?P>name), which is an extension to
1459     the Python syntax that PCRE uses for named parentheses (Perl does not provide
1460     named parentheses). We could rewrite the above example as follows:
1461 nigel 75 .sp
1462     (?P<pn> \e( ( (?>[^()]+) | (?P>pn) )* \e) )
1463     .sp
1464 nigel 63 This particular example pattern contains nested unlimited repeats, and so the
1465     use of atomic grouping for matching strings of non-parentheses is important
1466     when applying the pattern to strings that do not match. For example, when this
1467     pattern is applied to
1468 nigel 75 .sp
1469 nigel 63 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
1470 nigel 75 .sp
1471 nigel 63 it yields "no match" quickly. However, if atomic grouping is not used,
1472     the match runs for a very long time indeed because there are so many different
1473     ways the + and * repeats can carve up the subject, and all have to be tested
1474     before failure can be reported.
1475 nigel 75 .P
1476 nigel 63 At the end of a match, the values set for any capturing subpatterns are those
1477     from the outermost level of the recursion at which the subpattern value is set.
1478     If you want to obtain intermediate values, a callout function can be used (see
1479 nigel 75 the next section and the
1480 nigel 63 .\" HREF
1481 nigel 75 \fBpcrecallout\fP
1482 nigel 63 .\"
1483     documentation). If the pattern above is matched against
1484 nigel 75 .sp
1485 nigel 63 (ab(cd)ef)
1486 nigel 75 .sp
1487 nigel 63 the value for the capturing parentheses is "ef", which is the last value taken
1488     on at the top level. If additional parentheses are added, giving
1489 nigel 75 .sp
1490     \e( ( ( (?>[^()]+) | (?R) )* ) \e)
1491 nigel 63 ^ ^
1492     ^ ^
1493 nigel 75 .sp
1494 nigel 63 the string they capture is "ab(cd)ef", the contents of the top level
1495     parentheses. If there are more than 15 capturing parentheses in a pattern, PCRE
1496     has to obtain extra memory to store data during a recursion, which it does by
1497 nigel 75 using \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no
1498 nigel 63 memory can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
1499 nigel 75 .P
1500 nigel 63 Do not confuse the (?R) item with the condition (R), which tests for recursion.
1501     Consider this pattern, which matches text in angle brackets, allowing for
1502     arbitrary nesting. Only digits are allowed in nested brackets (that is, when
1503     recursing), whereas any characters are permitted at the outer level.
1504 nigel 75 .sp
1505     < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
1506     .sp
1507 nigel 63 In this pattern, (?(R) is the start of a conditional subpattern, with two
1508     different alternatives for the recursive and non-recursive cases. The (?R) item
1509     is the actual recursive call.
1510 nigel 75 .
1511     .
1512 nigel 63 .\" HTML <a name="subpatternsassubroutines"></a>
1513 nigel 75 .SH "SUBPATTERNS AS SUBROUTINES"
1514 nigel 63 .rs
1515     .sp
1516     If the syntax for a recursive subpattern reference (either by number or by
1517     name) is used outside the parentheses to which it refers, it operates like a
1518     subroutine in a programming language. An earlier example pointed out that the
1519     pattern
1520 nigel 75 .sp
1521     (sens|respons)e and \e1ibility
1522     .sp
1523 nigel 63 matches "sense and sensibility" and "response and responsibility", but not
1524     "sense and responsibility". If instead the pattern
1525 nigel 75 .sp
1526 nigel 63 (sens|respons)e and (?1)ibility
1527 nigel 75 .sp
1528 nigel 63 is used, it does match "sense and responsibility" as well as the other two
1529     strings. Such references must, however, follow the subpattern to which they
1530     refer.
1531 nigel 87 .P
1532     Like recursive subpatterns, a "subroutine" call is always treated as an atomic
1533     group. That is, once it has matched some of the subject string, it is never
1534     re-entered, even if it contains untried alternatives and there is a subsequent
1535     matching failure.
1536 nigel 75 .
1537     .
1538 nigel 63 .SH CALLOUTS
1539     .rs
1540     .sp
1541     Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
1542     code to be obeyed in the middle of matching a regular expression. This makes it
1543     possible, amongst other things, to extract different substrings that match the
1544     same pair of parentheses when there is a repetition.
1545 nigel 75 .P
1546 nigel 63 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
1547     code. The feature is called "callout". The caller of PCRE provides an external
1548 nigel 75 function by putting its entry point in the global variable \fIpcre_callout\fP.
1549 nigel 63 By default, this variable contains NULL, which disables all calling out.
1550 nigel 75 .P
1551 nigel 63 Within a regular expression, (?C) indicates the points at which the external
1552     function is to be called. If you want to identify different callout points, you
1553     can put a number less than 256 after the letter C. The default value is zero.
1554     For example, this pattern has two callout points:
1555 nigel 75 .sp
1556 nigel 63 (?C1)\dabc(?C2)def
1557 nigel 75 .sp
1558     If the PCRE_AUTO_CALLOUT flag is passed to \fBpcre_compile()\fP, callouts are
1559     automatically installed before each item in the pattern. They are all numbered
1560     255.
1561     .P
1562     During matching, when PCRE reaches a callout point (and \fIpcre_callout\fP is
1563 nigel 63 set), the external function is called. It is provided with the number of the
1564 nigel 75 callout, the position in the pattern, and, optionally, one item of data
1565     originally supplied by the caller of \fBpcre_exec()\fP. The callout function
1566     may cause matching to proceed, to backtrack, or to fail altogether. A complete
1567     description of the interface to the callout function is given in the
1568 nigel 63 .\" HREF
1569 nigel 75 \fBpcrecallout\fP
1570 nigel 63 .\"
1571     documentation.
1572 nigel 75 .P
1573 nigel 63 .in 0
1574 nigel 87 Last updated: 24 January 2006
1575 nigel 63 .br
1576 nigel 87 Copyright (c) 1997-2006 University of Cambridge.

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12