ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1412 - (show annotations)
Sun Dec 15 17:01:46 2013 UTC (5 years, 8 months ago) by ph10
File MIME type: text/html
File size: 139007 byte(s)
Final file tidies for 8.34.
1 <html>
2 <head>
3 <title>pcrepattern specification</title>
4 </head>
5 <body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB">
6 <h1>pcrepattern man page</h1>
7 <p>
8 Return to the <a href="index.html">PCRE index page</a>.
9 </p>
10 <p>
11 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 <br>
15 <ul>
16 <li><a name="TOC1" href="#SEC1">PCRE REGULAR EXPRESSION DETAILS</a>
17 <li><a name="TOC2" href="#SEC2">SPECIAL START-OF-PATTERN ITEMS</a>
18 <li><a name="TOC3" href="#SEC3">EBCDIC CHARACTER CODES</a>
19 <li><a name="TOC4" href="#SEC4">CHARACTERS AND METACHARACTERS</a>
20 <li><a name="TOC5" href="#SEC5">BACKSLASH</a>
21 <li><a name="TOC6" href="#SEC6">CIRCUMFLEX AND DOLLAR</a>
22 <li><a name="TOC7" href="#SEC7">FULL STOP (PERIOD, DOT) AND \N</a>
23 <li><a name="TOC8" href="#SEC8">MATCHING A SINGLE DATA UNIT</a>
25 <li><a name="TOC10" href="#SEC10">POSIX CHARACTER CLASSES</a>
27 <li><a name="TOC12" href="#SEC12">VERTICAL BAR</a>
28 <li><a name="TOC13" href="#SEC13">INTERNAL OPTION SETTING</a>
29 <li><a name="TOC14" href="#SEC14">SUBPATTERNS</a>
30 <li><a name="TOC15" href="#SEC15">DUPLICATE SUBPATTERN NUMBERS</a>
31 <li><a name="TOC16" href="#SEC16">NAMED SUBPATTERNS</a>
32 <li><a name="TOC17" href="#SEC17">REPETITION</a>
34 <li><a name="TOC19" href="#SEC19">BACK REFERENCES</a>
35 <li><a name="TOC20" href="#SEC20">ASSERTIONS</a>
36 <li><a name="TOC21" href="#SEC21">CONDITIONAL SUBPATTERNS</a>
37 <li><a name="TOC22" href="#SEC22">COMMENTS</a>
38 <li><a name="TOC23" href="#SEC23">RECURSIVE PATTERNS</a>
39 <li><a name="TOC24" href="#SEC24">SUBPATTERNS AS SUBROUTINES</a>
40 <li><a name="TOC25" href="#SEC25">ONIGURUMA SUBROUTINE SYNTAX</a>
41 <li><a name="TOC26" href="#SEC26">CALLOUTS</a>
42 <li><a name="TOC27" href="#SEC27">BACKTRACKING CONTROL</a>
43 <li><a name="TOC28" href="#SEC28">SEE ALSO</a>
44 <li><a name="TOC29" href="#SEC29">AUTHOR</a>
45 <li><a name="TOC30" href="#SEC30">REVISION</a>
46 </ul>
47 <br><a name="SEC1" href="#TOC1">PCRE REGULAR EXPRESSION DETAILS</a><br>
48 <P>
49 The syntax and semantics of the regular expressions that are supported by PCRE
50 are described in detail below. There is a quick-reference syntax summary in the
51 <a href="pcresyntax.html"><b>pcresyntax</b></a>
52 page. PCRE tries to match Perl syntax and semantics as closely as it can. PCRE
53 also supports some alternative regular expression syntax (which does not
54 conflict with the Perl syntax) in order to provide some compatibility with
55 regular expressions in Python, .NET, and Oniguruma.
56 </P>
57 <P>
58 Perl's regular expressions are described in its own documentation, and
59 regular expressions in general are covered in a number of books, some of which
60 have copious examples. Jeffrey Friedl's "Mastering Regular Expressions",
61 published by O'Reilly, covers regular expressions in great detail. This
62 description of PCRE's regular expressions is intended as reference material.
63 </P>
64 <P>
65 This document discusses the patterns that are supported by PCRE when one its
66 main matching functions, <b>pcre_exec()</b> (8-bit) or <b>pcre[16|32]_exec()</b>
67 (16- or 32-bit), is used. PCRE also has alternative matching functions,
68 <b>pcre_dfa_exec()</b> and <b>pcre[16|32_dfa_exec()</b>, which match using a
69 different algorithm that is not Perl-compatible. Some of the features discussed
70 below are not available when DFA matching is used. The advantages and
71 disadvantages of the alternative functions, and how they differ from the normal
72 functions, are discussed in the
73 <a href="pcrematching.html"><b>pcrematching</b></a>
74 page.
75 </P>
76 <br><a name="SEC2" href="#TOC1">SPECIAL START-OF-PATTERN ITEMS</a><br>
77 <P>
78 A number of options that can be passed to <b>pcre_compile()</b> can also be set
79 by special items at the start of a pattern. These are not Perl-compatible, but
80 are provided to make these options accessible to pattern writers who are not
81 able to change the program that processes the pattern. Any number of these
82 items may appear, but they must all be together right at the start of the
83 pattern string, and the letters must be in upper case.
84 </P>
85 <br><b>
86 UTF support
87 </b><br>
88 <P>
89 The original operation of PCRE was on strings of one-byte characters. However,
90 there is now also support for UTF-8 strings in the original library, an
91 extra library that supports 16-bit and UTF-16 character strings, and a
92 third library that supports 32-bit and UTF-32 character strings. To use these
93 features, PCRE must be built to include appropriate support. When using UTF
94 strings you must either call the compiling function with the PCRE_UTF8,
95 PCRE_UTF16, or PCRE_UTF32 option, or the pattern must start with one of
96 these special sequences:
97 <pre>
98 (*UTF8)
99 (*UTF16)
100 (*UTF32)
101 (*UTF)
102 </pre>
103 (*UTF) is a generic sequence that can be used with any of the libraries.
104 Starting a pattern with such a sequence is equivalent to setting the relevant
105 option. How setting a UTF mode affects pattern matching is mentioned in several
106 places below. There is also a summary of features in the
107 <a href="pcreunicode.html"><b>pcreunicode</b></a>
108 page.
109 </P>
110 <P>
111 Some applications that allow their users to supply patterns may wish to
112 restrict them to non-UTF data for security reasons. If the PCRE_NEVER_UTF
113 option is set at compile time, (*UTF) etc. are not allowed, and their
114 appearance causes an error.
115 </P>
116 <br><b>
117 Unicode property support
118 </b><br>
119 <P>
120 Another special sequence that may appear at the start of a pattern is (*UCP).
121 This has the same effect as setting the PCRE_UCP option: it causes sequences
122 such as \d and \w to use Unicode properties to determine character types,
123 instead of recognizing only characters with codes less than 128 via a lookup
124 table.
125 </P>
126 <br><b>
127 Disabling auto-possessification
128 </b><br>
129 <P>
130 If a pattern starts with (*NO_AUTO_POSSESS), it has the same effect as setting
131 the PCRE_NO_AUTO_POSSESS option at compile time. This stops PCRE from making
132 quantifiers possessive when what follows cannot match the repeated item. For
133 example, by default a+b is treated as a++b. For more details, see the
134 <a href="pcreapi.html"><b>pcreapi</b></a>
135 documentation.
136 </P>
137 <br><b>
138 Disabling start-up optimizations
139 </b><br>
140 <P>
141 If a pattern starts with (*NO_START_OPT), it has the same effect as setting the
142 PCRE_NO_START_OPTIMIZE option either at compile or matching time. This disables
143 several optimizations for quickly reaching "no match" results. For more
144 details, see the
145 <a href="pcreapi.html"><b>pcreapi</b></a>
146 documentation.
147 <a name="newlines"></a></P>
148 <br><b>
149 Newline conventions
150 </b><br>
151 <P>
152 PCRE supports five different conventions for indicating line breaks in
153 strings: a single CR (carriage return) character, a single LF (linefeed)
154 character, the two-character sequence CRLF, any of the three preceding, or any
155 Unicode newline sequence. The
156 <a href="pcreapi.html"><b>pcreapi</b></a>
157 page has
158 <a href="pcreapi.html#newlines">further discussion</a>
159 about newlines, and shows how to set the newline convention in the
160 <i>options</i> arguments for the compiling and matching functions.
161 </P>
162 <P>
163 It is also possible to specify a newline convention by starting a pattern
164 string with one of the following five sequences:
165 <pre>
166 (*CR) carriage return
167 (*LF) linefeed
168 (*CRLF) carriage return, followed by linefeed
169 (*ANYCRLF) any of the three above
170 (*ANY) all Unicode newline sequences
171 </pre>
172 These override the default and the options given to the compiling function. For
173 example, on a Unix system where LF is the default newline sequence, the pattern
174 <pre>
175 (*CR)a.b
176 </pre>
177 changes the convention to CR. That pattern matches "a\nb" because LF is no
178 longer a newline. If more than one of these settings is present, the last one
179 is used.
180 </P>
181 <P>
182 The newline convention affects where the circumflex and dollar assertions are
183 true. It also affects the interpretation of the dot metacharacter when
184 PCRE_DOTALL is not set, and the behaviour of \N. However, it does not affect
185 what the \R escape sequence matches. By default, this is any Unicode newline
186 sequence, for Perl compatibility. However, this can be changed; see the
187 description of \R in the section entitled
188 <a href="#newlineseq">"Newline sequences"</a>
189 below. A change of \R setting can be combined with a change of newline
190 convention.
191 </P>
192 <br><b>
193 Setting match and recursion limits
194 </b><br>
195 <P>
196 The caller of <b>pcre_exec()</b> can set a limit on the number of times the
197 internal <b>match()</b> function is called and on the maximum depth of
198 recursive calls. These facilities are provided to catch runaway matches that
199 are provoked by patterns with huge matching trees (a typical example is a
200 pattern with nested unlimited repeats) and to avoid running out of system stack
201 by too much recursion. When one of these limits is reached, <b>pcre_exec()</b>
202 gives an error return. The limits can also be set by items at the start of the
203 pattern of the form
204 <pre>
205 (*LIMIT_MATCH=d)
207 </pre>
208 where d is any number of decimal digits. However, the value of the setting must
209 be less than the value set (or defaulted) by the caller of <b>pcre_exec()</b>
210 for it to have any effect. In other words, the pattern writer can lower the
211 limits set by the programmer, but not raise them. If there is more than one
212 setting of one of these limits, the lower value is used.
213 </P>
214 <br><a name="SEC3" href="#TOC1">EBCDIC CHARACTER CODES</a><br>
215 <P>
216 PCRE can be compiled to run in an environment that uses EBCDIC as its character
217 code rather than ASCII or Unicode (typically a mainframe system). In the
218 sections below, character code values are ASCII or Unicode; in an EBCDIC
219 environment these characters may have different code values, and there are no
220 code points greater than 255.
221 </P>
222 <br><a name="SEC4" href="#TOC1">CHARACTERS AND METACHARACTERS</a><br>
223 <P>
224 A regular expression is a pattern that is matched against a subject string from
225 left to right. Most characters stand for themselves in a pattern, and match the
226 corresponding characters in the subject. As a trivial example, the pattern
227 <pre>
228 The quick brown fox
229 </pre>
230 matches a portion of a subject string that is identical to itself. When
231 caseless matching is specified (the PCRE_CASELESS option), letters are matched
232 independently of case. In a UTF mode, PCRE always understands the concept of
233 case for characters whose values are less than 128, so caseless matching is
234 always possible. For characters with higher values, the concept of case is
235 supported if PCRE is compiled with Unicode property support, but not otherwise.
236 If you want to use caseless matching for characters 128 and above, you must
237 ensure that PCRE is compiled with Unicode property support as well as with
238 UTF support.
239 </P>
240 <P>
241 The power of regular expressions comes from the ability to include alternatives
242 and repetitions in the pattern. These are encoded in the pattern by the use of
243 <i>metacharacters</i>, which do not stand for themselves but instead are
244 interpreted in some special way.
245 </P>
246 <P>
247 There are two different sets of metacharacters: those that are recognized
248 anywhere in the pattern except within square brackets, and those that are
249 recognized within square brackets. Outside square brackets, the metacharacters
250 are as follows:
251 <pre>
252 \ general escape character with several uses
253 ^ assert start of string (or line, in multiline mode)
254 $ assert end of string (or line, in multiline mode)
255 . match any character except newline (by default)
256 [ start character class definition
257 | start of alternative branch
258 ( start subpattern
259 ) end subpattern
260 ? extends the meaning of (
261 also 0 or 1 quantifier
262 also quantifier minimizer
263 * 0 or more quantifier
264 + 1 or more quantifier
265 also "possessive quantifier"
266 { start min/max quantifier
267 </pre>
268 Part of a pattern that is in square brackets is called a "character class". In
269 a character class the only metacharacters are:
270 <pre>
271 \ general escape character
272 ^ negate the class, but only if the first character
273 - indicates character range
274 [ POSIX character class (only if followed by POSIX syntax)
275 ] terminates the character class
276 </pre>
277 The following sections describe the use of each of the metacharacters.
278 </P>
279 <br><a name="SEC5" href="#TOC1">BACKSLASH</a><br>
280 <P>
281 The backslash character has several uses. Firstly, if it is followed by a
282 character that is not a number or a letter, it takes away any special meaning
283 that character may have. This use of backslash as an escape character applies
284 both inside and outside character classes.
285 </P>
286 <P>
287 For example, if you want to match a * character, you write \* in the pattern.
288 This escaping action applies whether or not the following character would
289 otherwise be interpreted as a metacharacter, so it is always safe to precede a
290 non-alphanumeric with backslash to specify that it stands for itself. In
291 particular, if you want to match a backslash, you write \\.
292 </P>
293 <P>
294 In a UTF mode, only ASCII numbers and letters have any special meaning after a
295 backslash. All other characters (in particular, those whose codepoints are
296 greater than 127) are treated as literals.
297 </P>
298 <P>
299 If a pattern is compiled with the PCRE_EXTENDED option, most white space in the
300 pattern (other than in a character class), and characters between a # outside a
301 character class and the next newline, inclusive, are ignored. An escaping
302 backslash can be used to include a white space or # character as part of the
303 pattern.
304 </P>
305 <P>
306 If you want to remove the special meaning from a sequence of characters, you
307 can do so by putting them between \Q and \E. This is different from Perl in
308 that $ and @ are handled as literals in \Q...\E sequences in PCRE, whereas in
309 Perl, $ and @ cause variable interpolation. Note the following examples:
310 <pre>
311 Pattern PCRE matches Perl matches
313 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
314 \Qabc\$xyz\E abc\$xyz abc\$xyz
315 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
316 </pre>
317 The \Q...\E sequence is recognized both inside and outside character classes.
318 An isolated \E that is not preceded by \Q is ignored. If \Q is not followed
319 by \E later in the pattern, the literal interpretation continues to the end of
320 the pattern (that is, \E is assumed at the end). If the isolated \Q is inside
321 a character class, this causes an error, because the character class is not
322 terminated.
323 <a name="digitsafterbackslash"></a></P>
324 <br><b>
325 Non-printing characters
326 </b><br>
327 <P>
328 A second use of backslash provides a way of encoding non-printing characters
329 in patterns in a visible manner. There is no restriction on the appearance of
330 non-printing characters, apart from the binary zero that terminates a pattern,
331 but when a pattern is being prepared by text editing, it is often easier to use
332 one of the following escape sequences than the binary character it represents:
333 <pre>
334 \a alarm, that is, the BEL character (hex 07)
335 \cx "control-x", where x is any ASCII character
336 \e escape (hex 1B)
337 \f form feed (hex 0C)
338 \n linefeed (hex 0A)
339 \r carriage return (hex 0D)
340 \t tab (hex 09)
341 \0dd character with octal code 0dd
342 \ddd character with octal code ddd, or back reference
343 \o{ddd..} character with octal code ddd..
344 \xhh character with hex code hh
345 \x{hhh..} character with hex code hhh.. (non-JavaScript mode)
346 \uhhhh character with hex code hhhh (JavaScript mode only)
347 </pre>
348 The precise effect of \cx on ASCII characters is as follows: if x is a lower
349 case letter, it is converted to upper case. Then bit 6 of the character (hex
350 40) is inverted. Thus \cA to \cZ become hex 01 to hex 1A (A is 41, Z is 5A),
351 but \c{ becomes hex 3B ({ is 7B), and \c; becomes hex 7B (; is 3B). If the
352 data item (byte or 16-bit value) following \c has a value greater than 127, a
353 compile-time error occurs. This locks out non-ASCII characters in all modes.
354 </P>
355 <P>
356 The \c facility was designed for use with ASCII characters, but with the
357 extension to Unicode it is even less useful than it once was. It is, however,
358 recognized when PCRE is compiled in EBCDIC mode, where data items are always
359 bytes. In this mode, all values are valid after \c. If the next character is a
360 lower case letter, it is converted to upper case. Then the 0xc0 bits of the
361 byte are inverted. Thus \cA becomes hex 01, as in ASCII (A is C1), but because
362 the EBCDIC letters are disjoint, \cZ becomes hex 29 (Z is E9), and other
363 characters also generate different values.
364 </P>
365 <P>
366 After \0 up to two further octal digits are read. If there are fewer than two
367 digits, just those that are present are used. Thus the sequence \0\x\07
368 specifies two binary zeros followed by a BEL character (code value 7). Make
369 sure you supply two digits after the initial zero if the pattern character that
370 follows is itself an octal digit.
371 </P>
372 <P>
373 The escape \o must be followed by a sequence of octal digits, enclosed in
374 braces. An error occurs if this is not the case. This escape is a recent
375 addition to Perl; it provides way of specifying character code points as octal
376 numbers greater than 0777, and it also allows octal numbers and back references
377 to be unambiguously specified.
378 </P>
379 <P>
380 For greater clarity and unambiguity, it is best to avoid following \ by a
381 digit greater than zero. Instead, use \o{} or \x{} to specify character
382 numbers, and \g{} to specify back references. The following paragraphs
383 describe the old, ambiguous syntax.
384 </P>
385 <P>
386 The handling of a backslash followed by a digit other than 0 is complicated,
387 and Perl has changed in recent releases, causing PCRE also to change. Outside a
388 character class, PCRE reads the digit and any following digits as a decimal
389 number. If the number is less than 8, or if there have been at least that many
390 previous capturing left parentheses in the expression, the entire sequence is
391 taken as a <i>back reference</i>. A description of how this works is given
392 <a href="#backreferences">later,</a>
393 following the discussion of
394 <a href="#subpattern">parenthesized subpatterns.</a>
395 </P>
396 <P>
397 Inside a character class, or if the decimal number following \ is greater than
398 7 and there have not been that many capturing subpatterns, PCRE handles \8 and
399 \9 as the literal characters "8" and "9", and otherwise re-reads up to three
400 octal digits following the backslash, using them to generate a data character.
401 Any subsequent digits stand for themselves. For example:
402 <pre>
403 \040 is another way of writing an ASCII space
404 \40 is the same, provided there are fewer than 40 previous capturing subpatterns
405 \7 is always a back reference
406 \11 might be a back reference, or another way of writing a tab
407 \011 is always a tab
408 \0113 is a tab followed by the character "3"
409 \113 might be a back reference, otherwise the character with octal code 113
410 \377 might be a back reference, otherwise the value 255 (decimal)
411 \81 is either a back reference, or the two characters "8" and "1"
412 </pre>
413 Note that octal values of 100 or greater that are specified using this syntax
414 must not be introduced by a leading zero, because no more than three octal
415 digits are ever read.
416 </P>
417 <P>
418 By default, after \x that is not followed by {, from zero to two hexadecimal
419 digits are read (letters can be in upper or lower case). Any number of
420 hexadecimal digits may appear between \x{ and }. If a character other than
421 a hexadecimal digit appears between \x{ and }, or if there is no terminating
422 }, an error occurs.
423 </P>
424 <P>
425 If the PCRE_JAVASCRIPT_COMPAT option is set, the interpretation of \x is
426 as just described only when it is followed by two hexadecimal digits.
427 Otherwise, it matches a literal "x" character. In JavaScript mode, support for
428 code points greater than 256 is provided by \u, which must be followed by
429 four hexadecimal digits; otherwise it matches a literal "u" character.
430 </P>
431 <P>
432 Characters whose value is less than 256 can be defined by either of the two
433 syntaxes for \x (or by \u in JavaScript mode). There is no difference in the
434 way they are handled. For example, \xdc is exactly the same as \x{dc} (or
435 \u00dc in JavaScript mode).
436 </P>
437 <br><b>
438 Constraints on character values
439 </b><br>
440 <P>
441 Characters that are specified using octal or hexadecimal numbers are
442 limited to certain values, as follows:
443 <pre>
444 8-bit non-UTF mode less than 0x100
445 8-bit UTF-8 mode less than 0x10ffff and a valid codepoint
446 16-bit non-UTF mode less than 0x10000
447 16-bit UTF-16 mode less than 0x10ffff and a valid codepoint
448 32-bit non-UTF mode less than 0x100000000
449 32-bit UTF-32 mode less than 0x10ffff and a valid codepoint
450 </pre>
451 Invalid Unicode codepoints are the range 0xd800 to 0xdfff (the so-called
452 "surrogate" codepoints), and 0xffef.
453 </P>
454 <br><b>
455 Escape sequences in character classes
456 </b><br>
457 <P>
458 All the sequences that define a single character value can be used both inside
459 and outside character classes. In addition, inside a character class, \b is
460 interpreted as the backspace character (hex 08).
461 </P>
462 <P>
463 \N is not allowed in a character class. \B, \R, and \X are not special
464 inside a character class. Like other unrecognized escape sequences, they are
465 treated as the literal characters "B", "R", and "X" by default, but cause an
466 error if the PCRE_EXTRA option is set. Outside a character class, these
467 sequences have different meanings.
468 </P>
469 <br><b>
470 Unsupported escape sequences
471 </b><br>
472 <P>
473 In Perl, the sequences \l, \L, \u, and \U are recognized by its string
474 handler and used to modify the case of following characters. By default, PCRE
475 does not support these escape sequences. However, if the PCRE_JAVASCRIPT_COMPAT
476 option is set, \U matches a "U" character, and \u can be used to define a
477 character by code point, as described in the previous section.
478 </P>
479 <br><b>
480 Absolute and relative back references
481 </b><br>
482 <P>
483 The sequence \g followed by an unsigned or a negative number, optionally
484 enclosed in braces, is an absolute or relative back reference. A named back
485 reference can be coded as \g{name}. Back references are discussed
486 <a href="#backreferences">later,</a>
487 following the discussion of
488 <a href="#subpattern">parenthesized subpatterns.</a>
489 </P>
490 <br><b>
491 Absolute and relative subroutine calls
492 </b><br>
493 <P>
494 For compatibility with Oniguruma, the non-Perl syntax \g followed by a name or
495 a number enclosed either in angle brackets or single quotes, is an alternative
496 syntax for referencing a subpattern as a "subroutine". Details are discussed
497 <a href="#onigurumasubroutines">later.</a>
498 Note that \g{...} (Perl syntax) and \g&#60;...&#62; (Oniguruma syntax) are <i>not</i>
499 synonymous. The former is a back reference; the latter is a
500 <a href="#subpatternsassubroutines">subroutine</a>
501 call.
502 <a name="genericchartypes"></a></P>
503 <br><b>
504 Generic character types
505 </b><br>
506 <P>
507 Another use of backslash is for specifying generic character types:
508 <pre>
509 \d any decimal digit
510 \D any character that is not a decimal digit
511 \h any horizontal white space character
512 \H any character that is not a horizontal white space character
513 \s any white space character
514 \S any character that is not a white space character
515 \v any vertical white space character
516 \V any character that is not a vertical white space character
517 \w any "word" character
518 \W any "non-word" character
519 </pre>
520 There is also the single sequence \N, which matches a non-newline character.
521 This is the same as
522 <a href="#fullstopdot">the "." metacharacter</a>
523 when PCRE_DOTALL is not set. Perl also uses \N to match characters by name;
524 PCRE does not support this.
525 </P>
526 <P>
527 Each pair of lower and upper case escape sequences partitions the complete set
528 of characters into two disjoint sets. Any given character matches one, and only
529 one, of each pair. The sequences can appear both inside and outside character
530 classes. They each match one character of the appropriate type. If the current
531 matching point is at the end of the subject string, all of them fail, because
532 there is no character to match.
533 </P>
534 <P>
535 For compatibility with Perl, \s did not used to match the VT character (code
536 11), which made it different from the the POSIX "space" class. However, Perl
537 added VT at release 5.18, and PCRE followed suit at release 8.34. The default
538 \s characters are now HT (9), LF (10), VT (11), FF (12), CR (13), and space
539 (32), which are defined as white space in the "C" locale. This list may vary if
540 locale-specific matching is taking place. For example, in some locales the
541 "non-breaking space" character (\xA0) is recognized as white space, and in
542 others the VT character is not.
543 </P>
544 <P>
545 A "word" character is an underscore or any character that is a letter or digit.
546 By default, the definition of letters and digits is controlled by PCRE's
547 low-valued character tables, and may vary if locale-specific matching is taking
548 place (see
549 <a href="pcreapi.html#localesupport">"Locale support"</a>
550 in the
551 <a href="pcreapi.html"><b>pcreapi</b></a>
552 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
553 or "french" in Windows, some character codes greater than 127 are used for
554 accented letters, and these are then matched by \w. The use of locales with
555 Unicode is discouraged.
556 </P>
557 <P>
558 By default, characters whose code points are greater than 127 never match \d,
559 \s, or \w, and always match \D, \S, and \W, although this may vary for
560 characters in the range 128-255 when locale-specific matching is happening.
561 These escape sequences retain their original meanings from before Unicode
562 support was available, mainly for efficiency reasons. If PCRE is compiled with
563 Unicode property support, and the PCRE_UCP option is set, the behaviour is
564 changed so that Unicode properties are used to determine character types, as
565 follows:
566 <pre>
567 \d any character that matches \p{Nd} (decimal digit)
568 \s any character that matches \p{Z} or \h or \v
569 \w any character that matches \p{L} or \p{N}, plus underscore
570 </pre>
571 The upper case escapes match the inverse sets of characters. Note that \d
572 matches only decimal digits, whereas \w matches any Unicode digit, as well as
573 any Unicode letter, and underscore. Note also that PCRE_UCP affects \b, and
574 \B because they are defined in terms of \w and \W. Matching these sequences
575 is noticeably slower when PCRE_UCP is set.
576 </P>
577 <P>
578 The sequences \h, \H, \v, and \V are features that were added to Perl at
579 release 5.10. In contrast to the other sequences, which match only ASCII
580 characters by default, these always match certain high-valued code points,
581 whether or not PCRE_UCP is set. The horizontal space characters are:
582 <pre>
583 U+0009 Horizontal tab (HT)
584 U+0020 Space
585 U+00A0 Non-break space
586 U+1680 Ogham space mark
587 U+180E Mongolian vowel separator
588 U+2000 En quad
589 U+2001 Em quad
590 U+2002 En space
591 U+2003 Em space
592 U+2004 Three-per-em space
593 U+2005 Four-per-em space
594 U+2006 Six-per-em space
595 U+2007 Figure space
596 U+2008 Punctuation space
597 U+2009 Thin space
598 U+200A Hair space
599 U+202F Narrow no-break space
600 U+205F Medium mathematical space
601 U+3000 Ideographic space
602 </pre>
603 The vertical space characters are:
604 <pre>
605 U+000A Linefeed (LF)
606 U+000B Vertical tab (VT)
607 U+000C Form feed (FF)
608 U+000D Carriage return (CR)
609 U+0085 Next line (NEL)
610 U+2028 Line separator
611 U+2029 Paragraph separator
612 </pre>
613 In 8-bit, non-UTF-8 mode, only the characters with codepoints less than 256 are
614 relevant.
615 <a name="newlineseq"></a></P>
616 <br><b>
617 Newline sequences
618 </b><br>
619 <P>
620 Outside a character class, by default, the escape sequence \R matches any
621 Unicode newline sequence. In 8-bit non-UTF-8 mode \R is equivalent to the
622 following:
623 <pre>
624 (?&#62;\r\n|\n|\x0b|\f|\r|\x85)
625 </pre>
626 This is an example of an "atomic group", details of which are given
627 <a href="#atomicgroup">below.</a>
628 This particular group matches either the two-character sequence CR followed by
629 LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
630 U+000B), FF (form feed, U+000C), CR (carriage return, U+000D), or NEL (next
631 line, U+0085). The two-character sequence is treated as a single unit that
632 cannot be split.
633 </P>
634 <P>
635 In other modes, two additional characters whose codepoints are greater than 255
636 are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
637 Unicode character property support is not needed for these characters to be
638 recognized.
639 </P>
640 <P>
641 It is possible to restrict \R to match only CR, LF, or CRLF (instead of the
642 complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
643 either at compile time or when the pattern is matched. (BSR is an abbrevation
644 for "backslash R".) This can be made the default when PCRE is built; if this is
645 the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
646 It is also possible to specify these settings by starting a pattern string with
647 one of the following sequences:
648 <pre>
649 (*BSR_ANYCRLF) CR, LF, or CRLF only
650 (*BSR_UNICODE) any Unicode newline sequence
651 </pre>
652 These override the default and the options given to the compiling function, but
653 they can themselves be overridden by options given to a matching function. Note
654 that these special settings, which are not Perl-compatible, are recognized only
655 at the very start of a pattern, and that they must be in upper case. If more
656 than one of them is present, the last one is used. They can be combined with a
657 change of newline convention; for example, a pattern can start with:
658 <pre>
660 </pre>
661 They can also be combined with the (*UTF8), (*UTF16), (*UTF32), (*UTF) or
662 (*UCP) special sequences. Inside a character class, \R is treated as an
663 unrecognized escape sequence, and so matches the letter "R" by default, but
664 causes an error if PCRE_EXTRA is set.
665 <a name="uniextseq"></a></P>
666 <br><b>
667 Unicode character properties
668 </b><br>
669 <P>
670 When PCRE is built with Unicode character property support, three additional
671 escape sequences that match characters with specific properties are available.
672 When in 8-bit non-UTF-8 mode, these sequences are of course limited to testing
673 characters whose codepoints are less than 256, but they do work in this mode.
674 The extra escape sequences are:
675 <pre>
676 \p{<i>xx</i>} a character with the <i>xx</i> property
677 \P{<i>xx</i>} a character without the <i>xx</i> property
678 \X a Unicode extended grapheme cluster
679 </pre>
680 The property names represented by <i>xx</i> above are limited to the Unicode
681 script names, the general category properties, "Any", which matches any
682 character (including newline), and some special PCRE properties (described
683 in the
684 <a href="#extraprops">next section).</a>
685 Other Perl properties such as "InMusicalSymbols" are not currently supported by
686 PCRE. Note that \P{Any} does not match any characters, so always causes a
687 match failure.
688 </P>
689 <P>
690 Sets of Unicode characters are defined as belonging to certain scripts. A
691 character from one of these sets can be matched using a script name. For
692 example:
693 <pre>
694 \p{Greek}
695 \P{Han}
696 </pre>
697 Those that are not part of an identified script are lumped together as
698 "Common". The current list of scripts is:
699 </P>
700 <P>
701 Arabic,
702 Armenian,
703 Avestan,
704 Balinese,
705 Bamum,
706 Batak,
707 Bengali,
708 Bopomofo,
709 Brahmi,
710 Braille,
711 Buginese,
712 Buhid,
713 Canadian_Aboriginal,
714 Carian,
715 Chakma,
716 Cham,
717 Cherokee,
718 Common,
719 Coptic,
720 Cuneiform,
721 Cypriot,
722 Cyrillic,
723 Deseret,
724 Devanagari,
725 Egyptian_Hieroglyphs,
726 Ethiopic,
727 Georgian,
728 Glagolitic,
729 Gothic,
730 Greek,
731 Gujarati,
732 Gurmukhi,
733 Han,
734 Hangul,
735 Hanunoo,
736 Hebrew,
737 Hiragana,
738 Imperial_Aramaic,
739 Inherited,
740 Inscriptional_Pahlavi,
741 Inscriptional_Parthian,
742 Javanese,
743 Kaithi,
744 Kannada,
745 Katakana,
746 Kayah_Li,
747 Kharoshthi,
748 Khmer,
749 Lao,
750 Latin,
751 Lepcha,
752 Limbu,
753 Linear_B,
754 Lisu,
755 Lycian,
756 Lydian,
757 Malayalam,
758 Mandaic,
759 Meetei_Mayek,
760 Meroitic_Cursive,
761 Meroitic_Hieroglyphs,
762 Miao,
763 Mongolian,
764 Myanmar,
765 New_Tai_Lue,
766 Nko,
767 Ogham,
768 Old_Italic,
769 Old_Persian,
770 Old_South_Arabian,
771 Old_Turkic,
772 Ol_Chiki,
773 Oriya,
774 Osmanya,
775 Phags_Pa,
776 Phoenician,
777 Rejang,
778 Runic,
779 Samaritan,
780 Saurashtra,
781 Sharada,
782 Shavian,
783 Sinhala,
784 Sora_Sompeng,
785 Sundanese,
786 Syloti_Nagri,
787 Syriac,
788 Tagalog,
789 Tagbanwa,
790 Tai_Le,
791 Tai_Tham,
792 Tai_Viet,
793 Takri,
794 Tamil,
795 Telugu,
796 Thaana,
797 Thai,
798 Tibetan,
799 Tifinagh,
800 Ugaritic,
801 Vai,
802 Yi.
803 </P>
804 <P>
805 Each character has exactly one Unicode general category property, specified by
806 a two-letter abbreviation. For compatibility with Perl, negation can be
807 specified by including a circumflex between the opening brace and the property
808 name. For example, \p{^Lu} is the same as \P{Lu}.
809 </P>
810 <P>
811 If only one letter is specified with \p or \P, it includes all the general
812 category properties that start with that letter. In this case, in the absence
813 of negation, the curly brackets in the escape sequence are optional; these two
814 examples have the same effect:
815 <pre>
816 \p{L}
817 \pL
818 </pre>
819 The following general category property codes are supported:
820 <pre>
821 C Other
822 Cc Control
823 Cf Format
824 Cn Unassigned
825 Co Private use
826 Cs Surrogate
828 L Letter
829 Ll Lower case letter
830 Lm Modifier letter
831 Lo Other letter
832 Lt Title case letter
833 Lu Upper case letter
835 M Mark
836 Mc Spacing mark
837 Me Enclosing mark
838 Mn Non-spacing mark
840 N Number
841 Nd Decimal number
842 Nl Letter number
843 No Other number
845 P Punctuation
846 Pc Connector punctuation
847 Pd Dash punctuation
848 Pe Close punctuation
849 Pf Final punctuation
850 Pi Initial punctuation
851 Po Other punctuation
852 Ps Open punctuation
854 S Symbol
855 Sc Currency symbol
856 Sk Modifier symbol
857 Sm Mathematical symbol
858 So Other symbol
860 Z Separator
861 Zl Line separator
862 Zp Paragraph separator
863 Zs Space separator
864 </pre>
865 The special property L& is also supported: it matches a character that has
866 the Lu, Ll, or Lt property, in other words, a letter that is not classified as
867 a modifier or "other".
868 </P>
869 <P>
870 The Cs (Surrogate) property applies only to characters in the range U+D800 to
871 U+DFFF. Such characters are not valid in Unicode strings and so
872 cannot be tested by PCRE, unless UTF validity checking has been turned off
873 (see the discussion of PCRE_NO_UTF8_CHECK, PCRE_NO_UTF16_CHECK and
874 PCRE_NO_UTF32_CHECK in the
875 <a href="pcreapi.html"><b>pcreapi</b></a>
876 page). Perl does not support the Cs property.
877 </P>
878 <P>
879 The long synonyms for property names that Perl supports (such as \p{Letter})
880 are not supported by PCRE, nor is it permitted to prefix any of these
881 properties with "Is".
882 </P>
883 <P>
884 No character that is in the Unicode table has the Cn (unassigned) property.
885 Instead, this property is assumed for any code point that is not in the
886 Unicode table.
887 </P>
888 <P>
889 Specifying caseless matching does not affect these escape sequences. For
890 example, \p{Lu} always matches only upper case letters. This is different from
891 the behaviour of current versions of Perl.
892 </P>
893 <P>
894 Matching characters by Unicode property is not fast, because PCRE has to do a
895 multistage table lookup in order to find a character's property. That is why
896 the traditional escape sequences such as \d and \w do not use Unicode
897 properties in PCRE by default, though you can make them do so by setting the
898 PCRE_UCP option or by starting the pattern with (*UCP).
899 </P>
900 <br><b>
901 Extended grapheme clusters
902 </b><br>
903 <P>
904 The \X escape matches any number of Unicode characters that form an "extended
905 grapheme cluster", and treats the sequence as an atomic group
906 <a href="#atomicgroup">(see below).</a>
907 Up to and including release 8.31, PCRE matched an earlier, simpler definition
908 that was equivalent to
909 <pre>
910 (?&#62;\PM\pM*)
911 </pre>
912 That is, it matched a character without the "mark" property, followed by zero
913 or more characters with the "mark" property. Characters with the "mark"
914 property are typically non-spacing accents that affect the preceding character.
915 </P>
916 <P>
917 This simple definition was extended in Unicode to include more complicated
918 kinds of composite character by giving each character a grapheme breaking
919 property, and creating rules that use these properties to define the boundaries
920 of extended grapheme clusters. In releases of PCRE later than 8.31, \X matches
921 one of these clusters.
922 </P>
923 <P>
924 \X always matches at least one character. Then it decides whether to add
925 additional characters according to the following rules for ending a cluster:
926 </P>
927 <P>
928 1. End at the end of the subject string.
929 </P>
930 <P>
931 2. Do not end between CR and LF; otherwise end after any control character.
932 </P>
933 <P>
934 3. Do not break Hangul (a Korean script) syllable sequences. Hangul characters
935 are of five types: L, V, T, LV, and LVT. An L character may be followed by an
936 L, V, LV, or LVT character; an LV or V character may be followed by a V or T
937 character; an LVT or T character may be follwed only by a T character.
938 </P>
939 <P>
940 4. Do not end before extending characters or spacing marks. Characters with
941 the "mark" property always have the "extend" grapheme breaking property.
942 </P>
943 <P>
944 5. Do not end after prepend characters.
945 </P>
946 <P>
947 6. Otherwise, end the cluster.
948 <a name="extraprops"></a></P>
949 <br><b>
950 PCRE's additional properties
951 </b><br>
952 <P>
953 As well as the standard Unicode properties described above, PCRE supports four
954 more that make it possible to convert traditional escape sequences such as \w
955 and \s to use Unicode properties. PCRE uses these non-standard, non-Perl
956 properties internally when PCRE_UCP is set. However, they may also be used
957 explicitly. These properties are:
958 <pre>
959 Xan Any alphanumeric character
960 Xps Any POSIX space character
961 Xsp Any Perl space character
962 Xwd Any Perl "word" character
963 </pre>
964 Xan matches characters that have either the L (letter) or the N (number)
965 property. Xps matches the characters tab, linefeed, vertical tab, form feed, or
966 carriage return, and any other character that has the Z (separator) property.
967 Xsp is the same as Xps; it used to exclude vertical tab, for Perl
968 compatibility, but Perl changed, and so PCRE followed at release 8.34. Xwd
969 matches the same characters as Xan, plus underscore.
970 </P>
971 <P>
972 There is another non-standard property, Xuc, which matches any character that
973 can be represented by a Universal Character Name in C++ and other programming
974 languages. These are the characters $, @, ` (grave accent), and all characters
975 with Unicode code points greater than or equal to U+00A0, except for the
976 surrogates U+D800 to U+DFFF. Note that most base (ASCII) characters are
977 excluded. (Universal Character Names are of the form \uHHHH or \UHHHHHHHH
978 where H is a hexadecimal digit. Note that the Xuc property does not match these
979 sequences but the characters that they represent.)
980 <a name="resetmatchstart"></a></P>
981 <br><b>
982 Resetting the match start
983 </b><br>
984 <P>
985 The escape sequence \K causes any previously matched characters not to be
986 included in the final matched sequence. For example, the pattern:
987 <pre>
988 foo\Kbar
989 </pre>
990 matches "foobar", but reports that it has matched "bar". This feature is
991 similar to a lookbehind assertion
992 <a href="#lookbehind">(described below).</a>
993 However, in this case, the part of the subject before the real match does not
994 have to be of fixed length, as lookbehind assertions do. The use of \K does
995 not interfere with the setting of
996 <a href="#subpattern">captured substrings.</a>
997 For example, when the pattern
998 <pre>
999 (foo)\Kbar
1000 </pre>
1001 matches "foobar", the first substring is still set to "foo".
1002 </P>
1003 <P>
1004 Perl documents that the use of \K within assertions is "not well defined". In
1005 PCRE, \K is acted upon when it occurs inside positive assertions, but is
1006 ignored in negative assertions.
1007 <a name="smallassertions"></a></P>
1008 <br><b>
1009 Simple assertions
1010 </b><br>
1011 <P>
1012 The final use of backslash is for certain simple assertions. An assertion
1013 specifies a condition that has to be met at a particular point in a match,
1014 without consuming any characters from the subject string. The use of
1015 subpatterns for more complicated assertions is described
1016 <a href="#bigassertions">below.</a>
1017 The backslashed assertions are:
1018 <pre>
1019 \b matches at a word boundary
1020 \B matches when not at a word boundary
1021 \A matches at the start of the subject
1022 \Z matches at the end of the subject
1023 also matches before a newline at the end of the subject
1024 \z matches only at the end of the subject
1025 \G matches at the first matching position in the subject
1026 </pre>
1027 Inside a character class, \b has a different meaning; it matches the backspace
1028 character. If any other of these assertions appears in a character class, by
1029 default it matches the corresponding literal character (for example, \B
1030 matches the letter B). However, if the PCRE_EXTRA option is set, an "invalid
1031 escape sequence" error is generated instead.
1032 </P>
1033 <P>
1034 A word boundary is a position in the subject string where the current character
1035 and the previous character do not both match \w or \W (i.e. one matches
1036 \w and the other matches \W), or the start or end of the string if the
1037 first or last character matches \w, respectively. In a UTF mode, the meanings
1038 of \w and \W can be changed by setting the PCRE_UCP option. When this is
1039 done, it also affects \b and \B. Neither PCRE nor Perl has a separate "start
1040 of word" or "end of word" metasequence. However, whatever follows \b normally
1041 determines which it is. For example, the fragment \ba matches "a" at the start
1042 of a word.
1043 </P>
1044 <P>
1045 The \A, \Z, and \z assertions differ from the traditional circumflex and
1046 dollar (described in the next section) in that they only ever match at the very
1047 start and end of the subject string, whatever options are set. Thus, they are
1048 independent of multiline mode. These three assertions are not affected by the
1049 PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
1050 circumflex and dollar metacharacters. However, if the <i>startoffset</i>
1051 argument of <b>pcre_exec()</b> is non-zero, indicating that matching is to start
1052 at a point other than the beginning of the subject, \A can never match. The
1053 difference between \Z and \z is that \Z matches before a newline at the end
1054 of the string as well as at the very end, whereas \z matches only at the end.
1055 </P>
1056 <P>
1057 The \G assertion is true only when the current matching position is at the
1058 start point of the match, as specified by the <i>startoffset</i> argument of
1059 <b>pcre_exec()</b>. It differs from \A when the value of <i>startoffset</i> is
1060 non-zero. By calling <b>pcre_exec()</b> multiple times with appropriate
1061 arguments, you can mimic Perl's /g option, and it is in this kind of
1062 implementation where \G can be useful.
1063 </P>
1064 <P>
1065 Note, however, that PCRE's interpretation of \G, as the start of the current
1066 match, is subtly different from Perl's, which defines it as the end of the
1067 previous match. In Perl, these can be different when the previously matched
1068 string was empty. Because PCRE does just one match at a time, it cannot
1069 reproduce this behaviour.
1070 </P>
1071 <P>
1072 If all the alternatives of a pattern begin with \G, the expression is anchored
1073 to the starting match position, and the "anchored" flag is set in the compiled
1074 regular expression.
1075 </P>
1076 <br><a name="SEC6" href="#TOC1">CIRCUMFLEX AND DOLLAR</a><br>
1077 <P>
1078 The circumflex and dollar metacharacters are zero-width assertions. That is,
1079 they test for a particular condition being true without consuming any
1080 characters from the subject string.
1081 </P>
1082 <P>
1083 Outside a character class, in the default matching mode, the circumflex
1084 character is an assertion that is true only if the current matching point is at
1085 the start of the subject string. If the <i>startoffset</i> argument of
1086 <b>pcre_exec()</b> is non-zero, circumflex can never match if the PCRE_MULTILINE
1087 option is unset. Inside a character class, circumflex has an entirely different
1088 meaning
1089 <a href="#characterclass">(see below).</a>
1090 </P>
1091 <P>
1092 Circumflex need not be the first character of the pattern if a number of
1093 alternatives are involved, but it should be the first thing in each alternative
1094 in which it appears if the pattern is ever to match that branch. If all
1095 possible alternatives start with a circumflex, that is, if the pattern is
1096 constrained to match only at the start of the subject, it is said to be an
1097 "anchored" pattern. (There are also other constructs that can cause a pattern
1098 to be anchored.)
1099 </P>
1100 <P>
1101 The dollar character is an assertion that is true only if the current matching
1102 point is at the end of the subject string, or immediately before a newline at
1103 the end of the string (by default). Note, however, that it does not actually
1104 match the newline. Dollar need not be the last character of the pattern if a
1105 number of alternatives are involved, but it should be the last item in any
1106 branch in which it appears. Dollar has no special meaning in a character class.
1107 </P>
1108 <P>
1109 The meaning of dollar can be changed so that it matches only at the very end of
1110 the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
1111 does not affect the \Z assertion.
1112 </P>
1113 <P>
1114 The meanings of the circumflex and dollar characters are changed if the
1115 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
1116 immediately after internal newlines as well as at the start of the subject
1117 string. It does not match after a newline that ends the string. A dollar
1118 matches before any newlines in the string, as well as at the very end, when
1119 PCRE_MULTILINE is set. When newline is specified as the two-character
1120 sequence CRLF, isolated CR and LF characters do not indicate newlines.
1121 </P>
1122 <P>
1123 For example, the pattern /^abc$/ matches the subject string "def\nabc" (where
1124 \n represents a newline) in multiline mode, but not otherwise. Consequently,
1125 patterns that are anchored in single line mode because all branches start with
1126 ^ are not anchored in multiline mode, and a match for circumflex is possible
1127 when the <i>startoffset</i> argument of <b>pcre_exec()</b> is non-zero. The
1128 PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
1129 </P>
1130 <P>
1131 Note that the sequences \A, \Z, and \z can be used to match the start and
1132 end of the subject in both modes, and if all branches of a pattern start with
1133 \A it is always anchored, whether or not PCRE_MULTILINE is set.
1134 <a name="fullstopdot"></a></P>
1135 <br><a name="SEC7" href="#TOC1">FULL STOP (PERIOD, DOT) AND \N</a><br>
1136 <P>
1137 Outside a character class, a dot in the pattern matches any one character in
1138 the subject string except (by default) a character that signifies the end of a
1139 line.
1140 </P>
1141 <P>
1142 When a line ending is defined as a single character, dot never matches that
1143 character; when the two-character sequence CRLF is used, dot does not match CR
1144 if it is immediately followed by LF, but otherwise it matches all characters
1145 (including isolated CRs and LFs). When any Unicode line endings are being
1146 recognized, dot does not match CR or LF or any of the other line ending
1147 characters.
1148 </P>
1149 <P>
1150 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
1151 option is set, a dot matches any one character, without exception. If the
1152 two-character sequence CRLF is present in the subject string, it takes two dots
1153 to match it.
1154 </P>
1155 <P>
1156 The handling of dot is entirely independent of the handling of circumflex and
1157 dollar, the only relationship being that they both involve newlines. Dot has no
1158 special meaning in a character class.
1159 </P>
1160 <P>
1161 The escape sequence \N behaves like a dot, except that it is not affected by
1162 the PCRE_DOTALL option. In other words, it matches any character except one
1163 that signifies the end of a line. Perl also uses \N to match characters by
1164 name; PCRE does not support this.
1165 </P>
1166 <br><a name="SEC8" href="#TOC1">MATCHING A SINGLE DATA UNIT</a><br>
1167 <P>
1168 Outside a character class, the escape sequence \C matches any one data unit,
1169 whether or not a UTF mode is set. In the 8-bit library, one data unit is one
1170 byte; in the 16-bit library it is a 16-bit unit; in the 32-bit library it is
1171 a 32-bit unit. Unlike a dot, \C always
1172 matches line-ending characters. The feature is provided in Perl in order to
1173 match individual bytes in UTF-8 mode, but it is unclear how it can usefully be
1174 used. Because \C breaks up characters into individual data units, matching one
1175 unit with \C in a UTF mode means that the rest of the string may start with a
1176 malformed UTF character. This has undefined results, because PCRE assumes that
1177 it is dealing with valid UTF strings (and by default it checks this at the
1178 start of processing unless the PCRE_NO_UTF8_CHECK, PCRE_NO_UTF16_CHECK or
1179 PCRE_NO_UTF32_CHECK option is used).
1180 </P>
1181 <P>
1182 PCRE does not allow \C to appear in lookbehind assertions
1183 <a href="#lookbehind">(described below)</a>
1184 in a UTF mode, because this would make it impossible to calculate the length of
1185 the lookbehind.
1186 </P>
1187 <P>
1188 In general, the \C escape sequence is best avoided. However, one
1189 way of using it that avoids the problem of malformed UTF characters is to use a
1190 lookahead to check the length of the next character, as in this pattern, which
1191 could be used with a UTF-8 string (ignore white space and line breaks):
1192 <pre>
1193 (?| (?=[\x00-\x7f])(\C) |
1194 (?=[\x80-\x{7ff}])(\C)(\C) |
1195 (?=[\x{800}-\x{ffff}])(\C)(\C)(\C) |
1196 (?=[\x{10000}-\x{1fffff}])(\C)(\C)(\C)(\C))
1197 </pre>
1198 A group that starts with (?| resets the capturing parentheses numbers in each
1199 alternative (see
1200 <a href="#dupsubpatternnumber">"Duplicate Subpattern Numbers"</a>
1201 below). The assertions at the start of each branch check the next UTF-8
1202 character for values whose encoding uses 1, 2, 3, or 4 bytes, respectively. The
1203 character's individual bytes are then captured by the appropriate number of
1204 groups.
1205 <a name="characterclass"></a></P>
1206 <br><a name="SEC9" href="#TOC1">SQUARE BRACKETS AND CHARACTER CLASSES</a><br>
1207 <P>
1208 An opening square bracket introduces a character class, terminated by a closing
1209 square bracket. A closing square bracket on its own is not special by default.
1210 However, if the PCRE_JAVASCRIPT_COMPAT option is set, a lone closing square
1211 bracket causes a compile-time error. If a closing square bracket is required as
1212 a member of the class, it should be the first data character in the class
1213 (after an initial circumflex, if present) or escaped with a backslash.
1214 </P>
1215 <P>
1216 A character class matches a single character in the subject. In a UTF mode, the
1217 character may be more than one data unit long. A matched character must be in
1218 the set of characters defined by the class, unless the first character in the
1219 class definition is a circumflex, in which case the subject character must not
1220 be in the set defined by the class. If a circumflex is actually required as a
1221 member of the class, ensure it is not the first character, or escape it with a
1222 backslash.
1223 </P>
1224 <P>
1225 For example, the character class [aeiou] matches any lower case vowel, while
1226 [^aeiou] matches any character that is not a lower case vowel. Note that a
1227 circumflex is just a convenient notation for specifying the characters that
1228 are in the class by enumerating those that are not. A class that starts with a
1229 circumflex is not an assertion; it still consumes a character from the subject
1230 string, and therefore it fails if the current pointer is at the end of the
1231 string.
1232 </P>
1233 <P>
1234 In UTF-8 (UTF-16, UTF-32) mode, characters with values greater than 255 (0xffff)
1235 can be included in a class as a literal string of data units, or by using the
1236 \x{ escaping mechanism.
1237 </P>
1238 <P>
1239 When caseless matching is set, any letters in a class represent both their
1240 upper case and lower case versions, so for example, a caseless [aeiou] matches
1241 "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
1242 caseful version would. In a UTF mode, PCRE always understands the concept of
1243 case for characters whose values are less than 128, so caseless matching is
1244 always possible. For characters with higher values, the concept of case is
1245 supported if PCRE is compiled with Unicode property support, but not otherwise.
1246 If you want to use caseless matching in a UTF mode for characters 128 and
1247 above, you must ensure that PCRE is compiled with Unicode property support as
1248 well as with UTF support.
1249 </P>
1250 <P>
1251 Characters that might indicate line breaks are never treated in any special way
1252 when matching character classes, whatever line-ending sequence is in use, and
1253 whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
1254 such as [^a] always matches one of these characters.
1255 </P>
1256 <P>
1257 The minus (hyphen) character can be used to specify a range of characters in a
1258 character class. For example, [d-m] matches any letter between d and m,
1259 inclusive. If a minus character is required in a class, it must be escaped with
1260 a backslash or appear in a position where it cannot be interpreted as
1261 indicating a range, typically as the first or last character in the class, or
1262 immediately after a range. For example, [b-d-z] matches letters in the range b
1263 to d, a hyphen character, or z.
1264 </P>
1265 <P>
1266 It is not possible to have the literal character "]" as the end character of a
1267 range. A pattern such as [W-]46] is interpreted as a class of two characters
1268 ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
1269 "-46]". However, if the "]" is escaped with a backslash it is interpreted as
1270 the end of range, so [W-\]46] is interpreted as a class containing a range
1271 followed by two other characters. The octal or hexadecimal representation of
1272 "]" can also be used to end a range.
1273 </P>
1274 <P>
1275 An error is generated if a POSIX character class (see below) or an escape
1276 sequence other than one that defines a single character appears at a point
1277 where a range ending character is expected. For example, [z-\xff] is valid,
1278 but [A-\d] and [A-[:digit:]] are not.
1279 </P>
1280 <P>
1281 Ranges operate in the collating sequence of character values. They can also be
1282 used for characters specified numerically, for example [\000-\037]. Ranges
1283 can include any characters that are valid for the current mode.
1284 </P>
1285 <P>
1286 If a range that includes letters is used when caseless matching is set, it
1287 matches the letters in either case. For example, [W-c] is equivalent to
1288 [][\\^_`wxyzabc], matched caselessly, and in a non-UTF mode, if character
1289 tables for a French locale are in use, [\xc8-\xcb] matches accented E
1290 characters in both cases. In UTF modes, PCRE supports the concept of case for
1291 characters with values greater than 128 only when it is compiled with Unicode
1292 property support.
1293 </P>
1294 <P>
1295 The character escape sequences \d, \D, \h, \H, \p, \P, \s, \S, \v,
1296 \V, \w, and \W may appear in a character class, and add the characters that
1297 they match to the class. For example, [\dABCDEF] matches any hexadecimal
1298 digit. In UTF modes, the PCRE_UCP option affects the meanings of \d, \s, \w
1299 and their upper case partners, just as it does when they appear outside a
1300 character class, as described in the section entitled
1301 <a href="#genericchartypes">"Generic character types"</a>
1302 above. The escape sequence \b has a different meaning inside a character
1303 class; it matches the backspace character. The sequences \B, \N, \R, and \X
1304 are not special inside a character class. Like any other unrecognized escape
1305 sequences, they are treated as the literal characters "B", "N", "R", and "X" by
1306 default, but cause an error if the PCRE_EXTRA option is set.
1307 </P>
1308 <P>
1309 A circumflex can conveniently be used with the upper case character types to
1310 specify a more restricted set of characters than the matching lower case type.
1311 For example, the class [^\W_] matches any letter or digit, but not underscore,
1312 whereas [\w] includes underscore. A positive character class should be read as
1313 "something OR something OR ..." and a negative class as "NOT something AND NOT
1314 something AND NOT ...".
1315 </P>
1316 <P>
1317 The only metacharacters that are recognized in character classes are backslash,
1318 hyphen (only where it can be interpreted as specifying a range), circumflex
1319 (only at the start), opening square bracket (only when it can be interpreted as
1320 introducing a POSIX class name, or for a special compatibility feature - see
1321 the next two sections), and the terminating closing square bracket. However,
1322 escaping other non-alphanumeric characters does no harm.
1323 </P>
1324 <br><a name="SEC10" href="#TOC1">POSIX CHARACTER CLASSES</a><br>
1325 <P>
1326 Perl supports the POSIX notation for character classes. This uses names
1327 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
1328 this notation. For example,
1329 <pre>
1330 [01[:alpha:]%]
1331 </pre>
1332 matches "0", "1", any alphabetic character, or "%". The supported class names
1333 are:
1334 <pre>
1335 alnum letters and digits
1336 alpha letters
1337 ascii character codes 0 - 127
1338 blank space or tab only
1339 cntrl control characters
1340 digit decimal digits (same as \d)
1341 graph printing characters, excluding space
1342 lower lower case letters
1343 print printing characters, including space
1344 punct printing characters, excluding letters and digits and space
1345 space white space (the same as \s from PCRE 8.34)
1346 upper upper case letters
1347 word "word" characters (same as \w)
1348 xdigit hexadecimal digits
1349 </pre>
1350 The default "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
1351 and space (32). If locale-specific matching is taking place, the list of space
1352 characters may be different; there may be fewer or more of them. "Space" used
1353 to be different to \s, which did not include VT, for Perl compatibility.
1354 However, Perl changed at release 5.18, and PCRE followed at release 8.34.
1355 "Space" and \s now match the same set of characters.
1356 </P>
1357 <P>
1358 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
1359 5.8. Another Perl extension is negation, which is indicated by a ^ character
1360 after the colon. For example,
1361 <pre>
1362 [12[:^digit:]]
1363 </pre>
1364 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
1365 syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
1366 supported, and an error is given if they are encountered.
1367 </P>
1368 <P>
1369 By default, characters with values greater than 128 do not match any of the
1370 POSIX character classes. However, if the PCRE_UCP option is passed to
1371 <b>pcre_compile()</b>, some of the classes are changed so that Unicode character
1372 properties are used. This is achieved by replacing certain POSIX classes by
1373 other sequences, as follows:
1374 <pre>
1375 [:alnum:] becomes \p{Xan}
1376 [:alpha:] becomes \p{L}
1377 [:blank:] becomes \h
1378 [:digit:] becomes \p{Nd}
1379 [:lower:] becomes \p{Ll}
1380 [:space:] becomes \p{Xps}
1381 [:upper:] becomes \p{Lu}
1382 [:word:] becomes \p{Xwd}
1383 </pre>
1384 Negated versions, such as [:^alpha:] use \P instead of \p. Three other POSIX
1385 classes are handled specially in UCP mode:
1386 </P>
1387 <P>
1388 [:graph:]
1389 This matches characters that have glyphs that mark the page when printed. In
1390 Unicode property terms, it matches all characters with the L, M, N, P, S, or Cf
1391 properties, except for:
1392 <pre>
1393 U+061C Arabic Letter Mark
1394 U+180E Mongolian Vowel Separator
1395 U+2066 - U+2069 Various "isolate"s
1397 </PRE>
1398 </P>
1399 <P>
1400 [:print:]
1401 This matches the same characters as [:graph:] plus space characters that are
1402 not controls, that is, characters with the Zs property.
1403 </P>
1404 <P>
1405 [:punct:]
1406 This matches all characters that have the Unicode P (punctuation) property,
1407 plus those characters whose code points are less than 128 that have the S
1408 (Symbol) property.
1409 </P>
1410 <P>
1411 The other POSIX classes are unchanged, and match only characters with code
1412 points less than 128.
1413 </P>
1414 <br><a name="SEC11" href="#TOC1">COMPATIBILITY FEATURE FOR WORD BOUNDARIES</a><br>
1415 <P>
1416 In the POSIX.2 compliant library that was included in 4.4BSD Unix, the ugly
1417 syntax [[:&#60;:]] and [[:&#62;:]] is used for matching "start of word" and "end of
1418 word". PCRE treats these items as follows:
1419 <pre>
1420 [[:&#60;:]] is converted to \b(?=\w)
1421 [[:&#62;:]] is converted to \b(?&#60;=\w)
1422 </pre>
1423 Only these exact character sequences are recognized. A sequence such as
1424 [a[:&#60;:]b] provokes error for an unrecognized POSIX class name. This support is
1425 not compatible with Perl. It is provided to help migrations from other
1426 environments, and is best not used in any new patterns. Note that \b matches
1427 at the start and the end of a word (see
1428 <a href="#smallassertions">"Simple assertions"</a>
1429 above), and in a Perl-style pattern the preceding or following character
1430 normally shows which is wanted, without the need for the assertions that are
1431 used above in order to give exactly the POSIX behaviour.
1432 </P>
1433 <br><a name="SEC12" href="#TOC1">VERTICAL BAR</a><br>
1434 <P>
1435 Vertical bar characters are used to separate alternative patterns. For example,
1436 the pattern
1437 <pre>
1438 gilbert|sullivan
1439 </pre>
1440 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
1441 and an empty alternative is permitted (matching the empty string). The matching
1442 process tries each alternative in turn, from left to right, and the first one
1443 that succeeds is used. If the alternatives are within a subpattern
1444 <a href="#subpattern">(defined below),</a>
1445 "succeeds" means matching the rest of the main pattern as well as the
1446 alternative in the subpattern.
1447 </P>
1448 <br><a name="SEC13" href="#TOC1">INTERNAL OPTION SETTING</a><br>
1449 <P>
1450 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1451 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1452 the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1453 The option letters are
1454 <pre>
1455 i for PCRE_CASELESS
1457 s for PCRE_DOTALL
1458 x for PCRE_EXTENDED
1459 </pre>
1460 For example, (?im) sets caseless, multiline matching. It is also possible to
1461 unset these options by preceding the letter with a hyphen, and a combined
1462 setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1463 PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1464 permitted. If a letter appears both before and after the hyphen, the option is
1465 unset.
1466 </P>
1467 <P>
1468 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1469 changed in the same way as the Perl-compatible options by using the characters
1470 J, U and X respectively.
1471 </P>
1472 <P>
1473 When one of these option changes occurs at top level (that is, not inside
1474 subpattern parentheses), the change applies to the remainder of the pattern
1475 that follows. If the change is placed right at the start of a pattern, PCRE
1476 extracts it into the global options (and it will therefore show up in data
1477 extracted by the <b>pcre_fullinfo()</b> function).
1478 </P>
1479 <P>
1480 An option change within a subpattern (see below for a description of
1481 subpatterns) affects only that part of the subpattern that follows it, so
1482 <pre>
1483 (a(?i)b)c
1484 </pre>
1485 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1486 By this means, options can be made to have different settings in different
1487 parts of the pattern. Any changes made in one alternative do carry on
1488 into subsequent branches within the same subpattern. For example,
1489 <pre>
1490 (a(?i)b|c)
1491 </pre>
1492 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1493 branch is abandoned before the option setting. This is because the effects of
1494 option settings happen at compile time. There would be some very weird
1495 behaviour otherwise.
1496 </P>
1497 <P>
1498 <b>Note:</b> There are other PCRE-specific options that can be set by the
1499 application when the compiling or matching functions are called. In some cases
1500 the pattern can contain special leading sequences such as (*CRLF) to override
1501 what the application has set or what has been defaulted. Details are given in
1502 the section entitled
1503 <a href="#newlineseq">"Newline sequences"</a>
1504 above. There are also the (*UTF8), (*UTF16),(*UTF32), and (*UCP) leading
1505 sequences that can be used to set UTF and Unicode property modes; they are
1506 equivalent to setting the PCRE_UTF8, PCRE_UTF16, PCRE_UTF32 and the PCRE_UCP
1507 options, respectively. The (*UTF) sequence is a generic version that can be
1508 used with any of the libraries. However, the application can set the
1509 PCRE_NEVER_UTF option, which locks out the use of the (*UTF) sequences.
1510 <a name="subpattern"></a></P>
1511 <br><a name="SEC14" href="#TOC1">SUBPATTERNS</a><br>
1512 <P>
1513 Subpatterns are delimited by parentheses (round brackets), which can be nested.
1514 Turning part of a pattern into a subpattern does two things:
1515 <br>
1516 <br>
1517 1. It localizes a set of alternatives. For example, the pattern
1518 <pre>
1519 cat(aract|erpillar|)
1520 </pre>
1521 matches "cataract", "caterpillar", or "cat". Without the parentheses, it would
1522 match "cataract", "erpillar" or an empty string.
1523 <br>
1524 <br>
1525 2. It sets up the subpattern as a capturing subpattern. This means that, when
1526 the whole pattern matches, that portion of the subject string that matched the
1527 subpattern is passed back to the caller via the <i>ovector</i> argument of the
1528 matching function. (This applies only to the traditional matching functions;
1529 the DFA matching functions do not support capturing.)
1530 </P>
1531 <P>
1532 Opening parentheses are counted from left to right (starting from 1) to obtain
1533 numbers for the capturing subpatterns. For example, if the string "the red
1534 king" is matched against the pattern
1535 <pre>
1536 the ((red|white) (king|queen))
1537 </pre>
1538 the captured substrings are "red king", "red", and "king", and are numbered 1,
1539 2, and 3, respectively.
1540 </P>
1541 <P>
1542 The fact that plain parentheses fulfil two functions is not always helpful.
1543 There are often times when a grouping subpattern is required without a
1544 capturing requirement. If an opening parenthesis is followed by a question mark
1545 and a colon, the subpattern does not do any capturing, and is not counted when
1546 computing the number of any subsequent capturing subpatterns. For example, if
1547 the string "the white queen" is matched against the pattern
1548 <pre>
1549 the ((?:red|white) (king|queen))
1550 </pre>
1551 the captured substrings are "white queen" and "queen", and are numbered 1 and
1552 2. The maximum number of capturing subpatterns is 65535.
1553 </P>
1554 <P>
1555 As a convenient shorthand, if any option settings are required at the start of
1556 a non-capturing subpattern, the option letters may appear between the "?" and
1557 the ":". Thus the two patterns
1558 <pre>
1559 (?i:saturday|sunday)
1560 (?:(?i)saturday|sunday)
1561 </pre>
1562 match exactly the same set of strings. Because alternative branches are tried
1563 from left to right, and options are not reset until the end of the subpattern
1564 is reached, an option setting in one branch does affect subsequent branches, so
1565 the above patterns match "SUNDAY" as well as "Saturday".
1566 <a name="dupsubpatternnumber"></a></P>
1567 <br><a name="SEC15" href="#TOC1">DUPLICATE SUBPATTERN NUMBERS</a><br>
1568 <P>
1569 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1570 the same numbers for its capturing parentheses. Such a subpattern starts with
1571 (?| and is itself a non-capturing subpattern. For example, consider this
1572 pattern:
1573 <pre>
1574 (?|(Sat)ur|(Sun))day
1575 </pre>
1576 Because the two alternatives are inside a (?| group, both sets of capturing
1577 parentheses are numbered one. Thus, when the pattern matches, you can look
1578 at captured substring number one, whichever alternative matched. This construct
1579 is useful when you want to capture part, but not all, of one of a number of
1580 alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1581 number is reset at the start of each branch. The numbers of any capturing
1582 parentheses that follow the subpattern start after the highest number used in
1583 any branch. The following example is taken from the Perl documentation. The
1584 numbers underneath show in which buffer the captured content will be stored.
1585 <pre>
1586 # before ---------------branch-reset----------- after
1587 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1588 # 1 2 2 3 2 3 4
1589 </pre>
1590 A back reference to a numbered subpattern uses the most recent value that is
1591 set for that number by any subpattern. The following pattern matches "abcabc"
1592 or "defdef":
1593 <pre>
1594 /(?|(abc)|(def))\1/
1595 </pre>
1596 In contrast, a subroutine call to a numbered subpattern always refers to the
1597 first one in the pattern with the given number. The following pattern matches
1598 "abcabc" or "defabc":
1599 <pre>
1600 /(?|(abc)|(def))(?1)/
1601 </pre>
1602 If a
1603 <a href="#conditions">condition test</a>
1604 for a subpattern's having matched refers to a non-unique number, the test is
1605 true if any of the subpatterns of that number have matched.
1606 </P>
1607 <P>
1608 An alternative approach to using this "branch reset" feature is to use
1609 duplicate named subpatterns, as described in the next section.
1610 </P>
1611 <br><a name="SEC16" href="#TOC1">NAMED SUBPATTERNS</a><br>
1612 <P>
1613 Identifying capturing parentheses by number is simple, but it can be very hard
1614 to keep track of the numbers in complicated regular expressions. Furthermore,
1615 if an expression is modified, the numbers may change. To help with this
1616 difficulty, PCRE supports the naming of subpatterns. This feature was not
1617 added to Perl until release 5.10. Python had the feature earlier, and PCRE
1618 introduced it at release 4.0, using the Python syntax. PCRE now supports both
1619 the Perl and the Python syntax. Perl allows identically numbered subpatterns to
1620 have different names, but PCRE does not.
1621 </P>
1622 <P>
1623 In PCRE, a subpattern can be named in one of three ways: (?&#60;name&#62;...) or
1624 (?'name'...) as in Perl, or (?P&#60;name&#62;...) as in Python. References to capturing
1625 parentheses from other parts of the pattern, such as
1626 <a href="#backreferences">back references,</a>
1627 <a href="#recursion">recursion,</a>
1628 and
1629 <a href="#conditions">conditions,</a>
1630 can be made by name as well as by number.
1631 </P>
1632 <P>
1633 Names consist of up to 32 alphanumeric characters and underscores, but must
1634 start with a non-digit. Named capturing parentheses are still allocated numbers
1635 as well as names, exactly as if the names were not present. The PCRE API
1636 provides function calls for extracting the name-to-number translation table
1637 from a compiled pattern. There is also a convenience function for extracting a
1638 captured substring by name.
1639 </P>
1640 <P>
1641 By default, a name must be unique within a pattern, but it is possible to relax
1642 this constraint by setting the PCRE_DUPNAMES option at compile time. (Duplicate
1643 names are also always permitted for subpatterns with the same number, set up as
1644 described in the previous section.) Duplicate names can be useful for patterns
1645 where only one instance of the named parentheses can match. Suppose you want to
1646 match the name of a weekday, either as a 3-letter abbreviation or as the full
1647 name, and in both cases you want to extract the abbreviation. This pattern
1648 (ignoring the line breaks) does the job:
1649 <pre>
1650 (?&#60;DN&#62;Mon|Fri|Sun)(?:day)?|
1651 (?&#60;DN&#62;Tue)(?:sday)?|
1652 (?&#60;DN&#62;Wed)(?:nesday)?|
1653 (?&#60;DN&#62;Thu)(?:rsday)?|
1654 (?&#60;DN&#62;Sat)(?:urday)?
1655 </pre>
1656 There are five capturing substrings, but only one is ever set after a match.
1657 (An alternative way of solving this problem is to use a "branch reset"
1658 subpattern, as described in the previous section.)
1659 </P>
1660 <P>
1661 The convenience function for extracting the data by name returns the substring
1662 for the first (and in this example, the only) subpattern of that name that
1663 matched. This saves searching to find which numbered subpattern it was.
1664 </P>
1665 <P>
1666 If you make a back reference to a non-unique named subpattern from elsewhere in
1667 the pattern, the subpatterns to which the name refers are checked in the order
1668 in which they appear in the overall pattern. The first one that is set is used
1669 for the reference. For example, this pattern matches both "foofoo" and
1670 "barbar" but not "foobar" or "barfoo":
1671 <pre>
1672 (?:(?&#60;n&#62;foo)|(?&#60;n&#62;bar))\k&#60;n&#62;
1674 </PRE>
1675 </P>
1676 <P>
1677 If you make a subroutine call to a non-unique named subpattern, the one that
1678 corresponds to the first occurrence of the name is used. In the absence of
1679 duplicate numbers (see the previous section) this is the one with the lowest
1680 number.
1681 </P>
1682 <P>
1683 If you use a named reference in a condition
1684 test (see the
1685 <a href="#conditions">section about conditions</a>
1686 below), either to check whether a subpattern has matched, or to check for
1687 recursion, all subpatterns with the same name are tested. If the condition is
1688 true for any one of them, the overall condition is true. This is the same
1689 behaviour as testing by number. For further details of the interfaces for
1690 handling named subpatterns, see the
1691 <a href="pcreapi.html"><b>pcreapi</b></a>
1692 documentation.
1693 </P>
1694 <P>
1695 <b>Warning:</b> You cannot use different names to distinguish between two
1696 subpatterns with the same number because PCRE uses only the numbers when
1697 matching. For this reason, an error is given at compile time if different names
1698 are given to subpatterns with the same number. However, you can always give the
1699 same name to subpatterns with the same number, even when PCRE_DUPNAMES is not
1700 set.
1701 </P>
1702 <br><a name="SEC17" href="#TOC1">REPETITION</a><br>
1703 <P>
1704 Repetition is specified by quantifiers, which can follow any of the following
1705 items:
1706 <pre>
1707 a literal data character
1708 the dot metacharacter
1709 the \C escape sequence
1710 the \X escape sequence
1711 the \R escape sequence
1712 an escape such as \d or \pL that matches a single character
1713 a character class
1714 a back reference (see next section)
1715 a parenthesized subpattern (including assertions)
1716 a subroutine call to a subpattern (recursive or otherwise)
1717 </pre>
1718 The general repetition quantifier specifies a minimum and maximum number of
1719 permitted matches, by giving the two numbers in curly brackets (braces),
1720 separated by a comma. The numbers must be less than 65536, and the first must
1721 be less than or equal to the second. For example:
1722 <pre>
1723 z{2,4}
1724 </pre>
1725 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1726 character. If the second number is omitted, but the comma is present, there is
1727 no upper limit; if the second number and the comma are both omitted, the
1728 quantifier specifies an exact number of required matches. Thus
1729 <pre>
1730 [aeiou]{3,}
1731 </pre>
1732 matches at least 3 successive vowels, but may match many more, while
1733 <pre>
1734 \d{8}
1735 </pre>
1736 matches exactly 8 digits. An opening curly bracket that appears in a position
1737 where a quantifier is not allowed, or one that does not match the syntax of a
1738 quantifier, is taken as a literal character. For example, {,6} is not a
1739 quantifier, but a literal string of four characters.
1740 </P>
1741 <P>
1742 In UTF modes, quantifiers apply to characters rather than to individual data
1743 units. Thus, for example, \x{100}{2} matches two characters, each of
1744 which is represented by a two-byte sequence in a UTF-8 string. Similarly,
1745 \X{3} matches three Unicode extended grapheme clusters, each of which may be
1746 several data units long (and they may be of different lengths).
1747 </P>
1748 <P>
1749 The quantifier {0} is permitted, causing the expression to behave as if the
1750 previous item and the quantifier were not present. This may be useful for
1751 subpatterns that are referenced as
1752 <a href="#subpatternsassubroutines">subroutines</a>
1753 from elsewhere in the pattern (but see also the section entitled
1754 <a href="#subdefine">"Defining subpatterns for use by reference only"</a>
1755 below). Items other than subpatterns that have a {0} quantifier are omitted
1756 from the compiled pattern.
1757 </P>
1758 <P>
1759 For convenience, the three most common quantifiers have single-character
1760 abbreviations:
1761 <pre>
1762 * is equivalent to {0,}
1763 + is equivalent to {1,}
1764 ? is equivalent to {0,1}
1765 </pre>
1766 It is possible to construct infinite loops by following a subpattern that can
1767 match no characters with a quantifier that has no upper limit, for example:
1768 <pre>
1769 (a?)*
1770 </pre>
1771 Earlier versions of Perl and PCRE used to give an error at compile time for
1772 such patterns. However, because there are cases where this can be useful, such
1773 patterns are now accepted, but if any repetition of the subpattern does in fact
1774 match no characters, the loop is forcibly broken.
1775 </P>
1776 <P>
1777 By default, the quantifiers are "greedy", that is, they match as much as
1778 possible (up to the maximum number of permitted times), without causing the
1779 rest of the pattern to fail. The classic example of where this gives problems
1780 is in trying to match comments in C programs. These appear between /* and */
1781 and within the comment, individual * and / characters may appear. An attempt to
1782 match C comments by applying the pattern
1783 <pre>
1784 /\*.*\*/
1785 </pre>
1786 to the string
1787 <pre>
1788 /* first comment */ not comment /* second comment */
1789 </pre>
1790 fails, because it matches the entire string owing to the greediness of the .*
1791 item.
1792 </P>
1793 <P>
1794 However, if a quantifier is followed by a question mark, it ceases to be
1795 greedy, and instead matches the minimum number of times possible, so the
1796 pattern
1797 <pre>
1798 /\*.*?\*/
1799 </pre>
1800 does the right thing with the C comments. The meaning of the various
1801 quantifiers is not otherwise changed, just the preferred number of matches.
1802 Do not confuse this use of question mark with its use as a quantifier in its
1803 own right. Because it has two uses, it can sometimes appear doubled, as in
1804 <pre>
1805 \d??\d
1806 </pre>
1807 which matches one digit by preference, but can match two if that is the only
1808 way the rest of the pattern matches.
1809 </P>
1810 <P>
1811 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1812 the quantifiers are not greedy by default, but individual ones can be made
1813 greedy by following them with a question mark. In other words, it inverts the
1814 default behaviour.
1815 </P>
1816 <P>
1817 When a parenthesized subpattern is quantified with a minimum repeat count that
1818 is greater than 1 or with a limited maximum, more memory is required for the
1819 compiled pattern, in proportion to the size of the minimum or maximum.
1820 </P>
1821 <P>
1822 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1823 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1824 implicitly anchored, because whatever follows will be tried against every
1825 character position in the subject string, so there is no point in retrying the
1826 overall match at any position after the first. PCRE normally treats such a
1827 pattern as though it were preceded by \A.
1828 </P>
1829 <P>
1830 In cases where it is known that the subject string contains no newlines, it is
1831 worth setting PCRE_DOTALL in order to obtain this optimization, or
1832 alternatively using ^ to indicate anchoring explicitly.
1833 </P>
1834 <P>
1835 However, there are some cases where the optimization cannot be used. When .*
1836 is inside capturing parentheses that are the subject of a back reference
1837 elsewhere in the pattern, a match at the start may fail where a later one
1838 succeeds. Consider, for example:
1839 <pre>
1840 (.*)abc\1
1841 </pre>
1842 If the subject is "xyz123abc123" the match point is the fourth character. For
1843 this reason, such a pattern is not implicitly anchored.
1844 </P>
1845 <P>
1846 Another case where implicit anchoring is not applied is when the leading .* is
1847 inside an atomic group. Once again, a match at the start may fail where a later
1848 one succeeds. Consider this pattern:
1849 <pre>
1850 (?&#62;.*?a)b
1851 </pre>
1852 It matches "ab" in the subject "aab". The use of the backtracking control verbs
1853 (*PRUNE) and (*SKIP) also disable this optimization.
1854 </P>
1855 <P>
1856 When a capturing subpattern is repeated, the value captured is the substring
1857 that matched the final iteration. For example, after
1858 <pre>
1859 (tweedle[dume]{3}\s*)+
1860 </pre>
1861 has matched "tweedledum tweedledee" the value of the captured substring is
1862 "tweedledee". However, if there are nested capturing subpatterns, the
1863 corresponding captured values may have been set in previous iterations. For
1864 example, after
1865 <pre>
1866 /(a|(b))+/
1867 </pre>
1868 matches "aba" the value of the second captured substring is "b".
1869 <a name="atomicgroup"></a></P>
1870 <br><a name="SEC18" href="#TOC1">ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS</a><br>
1871 <P>
1872 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1873 repetition, failure of what follows normally causes the repeated item to be
1874 re-evaluated to see if a different number of repeats allows the rest of the
1875 pattern to match. Sometimes it is useful to prevent this, either to change the
1876 nature of the match, or to cause it fail earlier than it otherwise might, when
1877 the author of the pattern knows there is no point in carrying on.
1878 </P>
1879 <P>
1880 Consider, for example, the pattern \d+foo when applied to the subject line
1881 <pre>
1882 123456bar
1883 </pre>
1884 After matching all 6 digits and then failing to match "foo", the normal
1885 action of the matcher is to try again with only 5 digits matching the \d+
1886 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1887 (a term taken from Jeffrey Friedl's book) provides the means for specifying
1888 that once a subpattern has matched, it is not to be re-evaluated in this way.
1889 </P>
1890 <P>
1891 If we use atomic grouping for the previous example, the matcher gives up
1892 immediately on failing to match "foo" the first time. The notation is a kind of
1893 special parenthesis, starting with (?&#62; as in this example:
1894 <pre>
1895 (?&#62;\d+)foo
1896 </pre>
1897 This kind of parenthesis "locks up" the part of the pattern it contains once
1898 it has matched, and a failure further into the pattern is prevented from
1899 backtracking into it. Backtracking past it to previous items, however, works as
1900 normal.
1901 </P>
1902 <P>
1903 An alternative description is that a subpattern of this type matches the string
1904 of characters that an identical standalone pattern would match, if anchored at
1905 the current point in the subject string.
1906 </P>
1907 <P>
1908 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1909 the above example can be thought of as a maximizing repeat that must swallow
1910 everything it can. So, while both \d+ and \d+? are prepared to adjust the
1911 number of digits they match in order to make the rest of the pattern match,
1912 (?&#62;\d+) can only match an entire sequence of digits.
1913 </P>
1914 <P>
1915 Atomic groups in general can of course contain arbitrarily complicated
1916 subpatterns, and can be nested. However, when the subpattern for an atomic
1917 group is just a single repeated item, as in the example above, a simpler
1918 notation, called a "possessive quantifier" can be used. This consists of an
1919 additional + character following a quantifier. Using this notation, the
1920 previous example can be rewritten as
1921 <pre>
1922 \d++foo
1923 </pre>
1924 Note that a possessive quantifier can be used with an entire group, for
1925 example:
1926 <pre>
1927 (abc|xyz){2,3}+
1928 </pre>
1929 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1930 option is ignored. They are a convenient notation for the simpler forms of
1931 atomic group. However, there is no difference in the meaning of a possessive
1932 quantifier and the equivalent atomic group, though there may be a performance
1933 difference; possessive quantifiers should be slightly faster.
1934 </P>
1935 <P>
1936 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1937 Jeffrey Friedl originated the idea (and the name) in the first edition of his
1938 book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1939 package, and PCRE copied it from there. It ultimately found its way into Perl
1940 at release 5.10.
1941 </P>
1942 <P>
1943 PCRE has an optimization that automatically "possessifies" certain simple
1944 pattern constructs. For example, the sequence A+B is treated as A++B because
1945 there is no point in backtracking into a sequence of A's when B must follow.
1946 </P>
1947 <P>
1948 When a pattern contains an unlimited repeat inside a subpattern that can itself
1949 be repeated an unlimited number of times, the use of an atomic group is the
1950 only way to avoid some failing matches taking a very long time indeed. The
1951 pattern
1952 <pre>
1953 (\D+|&#60;\d+&#62;)*[!?]
1954 </pre>
1955 matches an unlimited number of substrings that either consist of non-digits, or
1956 digits enclosed in &#60;&#62;, followed by either ! or ?. When it matches, it runs
1957 quickly. However, if it is applied to
1958 <pre>
1959 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1960 </pre>
1961 it takes a long time before reporting failure. This is because the string can
1962 be divided between the internal \D+ repeat and the external * repeat in a
1963 large number of ways, and all have to be tried. (The example uses [!?] rather
1964 than a single character at the end, because both PCRE and Perl have an
1965 optimization that allows for fast failure when a single character is used. They
1966 remember the last single character that is required for a match, and fail early
1967 if it is not present in the string.) If the pattern is changed so that it uses
1968 an atomic group, like this:
1969 <pre>
1970 ((?&#62;\D+)|&#60;\d+&#62;)*[!?]
1971 </pre>
1972 sequences of non-digits cannot be broken, and failure happens quickly.
1973 <a name="backreferences"></a></P>
1974 <br><a name="SEC19" href="#TOC1">BACK REFERENCES</a><br>
1975 <P>
1976 Outside a character class, a backslash followed by a digit greater than 0 (and
1977 possibly further digits) is a back reference to a capturing subpattern earlier
1978 (that is, to its left) in the pattern, provided there have been that many
1979 previous capturing left parentheses.
1980 </P>
1981 <P>
1982 However, if the decimal number following the backslash is less than 10, it is
1983 always taken as a back reference, and causes an error only if there are not
1984 that many capturing left parentheses in the entire pattern. In other words, the
1985 parentheses that are referenced need not be to the left of the reference for
1986 numbers less than 10. A "forward back reference" of this type can make sense
1987 when a repetition is involved and the subpattern to the right has participated
1988 in an earlier iteration.
1989 </P>
1990 <P>
1991 It is not possible to have a numerical "forward back reference" to a subpattern
1992 whose number is 10 or more using this syntax because a sequence such as \50 is
1993 interpreted as a character defined in octal. See the subsection entitled
1994 "Non-printing characters"
1995 <a href="#digitsafterbackslash">above</a>
1996 for further details of the handling of digits following a backslash. There is
1997 no such problem when named parentheses are used. A back reference to any
1998 subpattern is possible using named parentheses (see below).
1999 </P>
2000 <P>
2001 Another way of avoiding the ambiguity inherent in the use of digits following a
2002 backslash is to use the \g escape sequence. This escape must be followed by an
2003 unsigned number or a negative number, optionally enclosed in braces. These
2004 examples are all identical:
2005 <pre>
2006 (ring), \1
2007 (ring), \g1
2008 (ring), \g{1}
2009 </pre>
2010 An unsigned number specifies an absolute reference without the ambiguity that
2011 is present in the older syntax. It is also useful when literal digits follow
2012 the reference. A negative number is a relative reference. Consider this
2013 example:
2014 <pre>
2015 (abc(def)ghi)\g{-1}
2016 </pre>
2017 The sequence \g{-1} is a reference to the most recently started capturing
2018 subpattern before \g, that is, is it equivalent to \2 in this example.
2019 Similarly, \g{-2} would be equivalent to \1. The use of relative references
2020 can be helpful in long patterns, and also in patterns that are created by
2021 joining together fragments that contain references within themselves.
2022 </P>
2023 <P>
2024 A back reference matches whatever actually matched the capturing subpattern in
2025 the current subject string, rather than anything matching the subpattern
2026 itself (see
2027 <a href="#subpatternsassubroutines">"Subpatterns as subroutines"</a>
2028 below for a way of doing that). So the pattern
2029 <pre>
2030 (sens|respons)e and \1ibility
2031 </pre>
2032 matches "sense and sensibility" and "response and responsibility", but not
2033 "sense and responsibility". If caseful matching is in force at the time of the
2034 back reference, the case of letters is relevant. For example,
2035 <pre>
2036 ((?i)rah)\s+\1
2037 </pre>
2038 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
2039 capturing subpattern is matched caselessly.
2040 </P>
2041 <P>
2042 There are several different ways of writing back references to named
2043 subpatterns. The .NET syntax \k{name} and the Perl syntax \k&#60;name&#62; or
2044 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
2045 back reference syntax, in which \g can be used for both numeric and named
2046 references, is also supported. We could rewrite the above example in any of
2047 the following ways:
2048 <pre>
2049 (?&#60;p1&#62;(?i)rah)\s+\k&#60;p1&#62;
2050 (?'p1'(?i)rah)\s+\k{p1}
2051 (?P&#60;p1&#62;(?i)rah)\s+(?P=p1)
2052 (?&#60;p1&#62;(?i)rah)\s+\g{p1}
2053 </pre>
2054 A subpattern that is referenced by name may appear in the pattern before or
2055 after the reference.
2056 </P>
2057 <P>
2058 There may be more than one back reference to the same subpattern. If a
2059 subpattern has not actually been used in a particular match, any back
2060 references to it always fail by default. For example, the pattern
2061 <pre>
2062 (a|(bc))\2
2063 </pre>
2064 always fails if it starts to match "a" rather than "bc". However, if the
2065 PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back reference to an
2066 unset value matches an empty string.
2067 </P>
2068 <P>
2069 Because there may be many capturing parentheses in a pattern, all digits
2070 following a backslash are taken as part of a potential back reference number.
2071 If the pattern continues with a digit character, some delimiter must be used to
2072 terminate the back reference. If the PCRE_EXTENDED option is set, this can be
2073 white space. Otherwise, the \g{ syntax or an empty comment (see
2074 <a href="#comments">"Comments"</a>
2075 below) can be used.
2076 </P>
2077 <br><b>
2078 Recursive back references
2079 </b><br>
2080 <P>
2081 A back reference that occurs inside the parentheses to which it refers fails
2082 when the subpattern is first used, so, for example, (a\1) never matches.
2083 However, such references can be useful inside repeated subpatterns. For
2084 example, the pattern
2085 <pre>
2086 (a|b\1)+
2087 </pre>
2088 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
2089 the subpattern, the back reference matches the character string corresponding
2090 to the previous iteration. In order for this to work, the pattern must be such
2091 that the first iteration does not need to match the back reference. This can be
2092 done using alternation, as in the example above, or by a quantifier with a
2093 minimum of zero.
2094 </P>
2095 <P>
2096 Back references of this type cause the group that they reference to be treated
2097 as an
2098 <a href="#atomicgroup">atomic group.</a>
2099 Once the whole group has been matched, a subsequent matching failure cannot
2100 cause backtracking into the middle of the group.
2101 <a name="bigassertions"></a></P>
2102 <br><a name="SEC20" href="#TOC1">ASSERTIONS</a><br>
2103 <P>
2104 An assertion is a test on the characters following or preceding the current
2105 matching point that does not actually consume any characters. The simple
2106 assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are described
2107 <a href="#smallassertions">above.</a>
2108 </P>
2109 <P>
2110 More complicated assertions are coded as subpatterns. There are two kinds:
2111 those that look ahead of the current position in the subject string, and those
2112 that look behind it. An assertion subpattern is matched in the normal way,
2113 except that it does not cause the current matching position to be changed.
2114 </P>
2115 <P>
2116 Assertion subpatterns are not capturing subpatterns. If such an assertion
2117 contains capturing subpatterns within it, these are counted for the purposes of
2118 numbering the capturing subpatterns in the whole pattern. However, substring
2119 capturing is carried out only for positive assertions. (Perl sometimes, but not
2120 always, does do capturing in negative assertions.)
2121 </P>
2122 <P>
2123 For compatibility with Perl, assertion subpatterns may be repeated; though
2124 it makes no sense to assert the same thing several times, the side effect of
2125 capturing parentheses may occasionally be useful. In practice, there only three
2126 cases:
2127 <br>
2128 <br>
2129 (1) If the quantifier is {0}, the assertion is never obeyed during matching.
2130 However, it may contain internal capturing parenthesized groups that are called
2131 from elsewhere via the
2132 <a href="#subpatternsassubroutines">subroutine mechanism.</a>
2133 <br>
2134 <br>
2135 (2) If quantifier is {0,n} where n is greater than zero, it is treated as if it
2136 were {0,1}. At run time, the rest of the pattern match is tried with and
2137 without the assertion, the order depending on the greediness of the quantifier.
2138 <br>
2139 <br>
2140 (3) If the minimum repetition is greater than zero, the quantifier is ignored.
2141 The assertion is obeyed just once when encountered during matching.
2142 </P>
2143 <br><b>
2144 Lookahead assertions
2145 </b><br>
2146 <P>
2147 Lookahead assertions start with (?= for positive assertions and (?! for
2148 negative assertions. For example,
2149 <pre>
2150 \w+(?=;)
2151 </pre>
2152 matches a word followed by a semicolon, but does not include the semicolon in
2153 the match, and
2154 <pre>
2155 foo(?!bar)
2156 </pre>
2157 matches any occurrence of "foo" that is not followed by "bar". Note that the
2158 apparently similar pattern
2159 <pre>
2160 (?!foo)bar
2161 </pre>
2162 does not find an occurrence of "bar" that is preceded by something other than
2163 "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
2164 (?!foo) is always true when the next three characters are "bar". A
2165 lookbehind assertion is needed to achieve the other effect.
2166 </P>
2167 <P>
2168 If you want to force a matching failure at some point in a pattern, the most
2169 convenient way to do it is with (?!) because an empty string always matches, so
2170 an assertion that requires there not to be an empty string must always fail.
2171 The backtracking control verb (*FAIL) or (*F) is a synonym for (?!).
2172 <a name="lookbehind"></a></P>
2173 <br><b>
2174 Lookbehind assertions
2175 </b><br>
2176 <P>
2177 Lookbehind assertions start with (?&#60;= for positive assertions and (?&#60;! for
2178 negative assertions. For example,
2179 <pre>
2180 (?&#60;!foo)bar
2181 </pre>
2182 does find an occurrence of "bar" that is not preceded by "foo". The contents of
2183 a lookbehind assertion are restricted such that all the strings it matches must
2184 have a fixed length. However, if there are several top-level alternatives, they
2185 do not all have to have the same fixed length. Thus
2186 <pre>
2187 (?&#60;=bullock|donkey)
2188 </pre>
2189 is permitted, but
2190 <pre>
2191 (?&#60;!dogs?|cats?)
2192 </pre>
2193 causes an error at compile time. Branches that match different length strings
2194 are permitted only at the top level of a lookbehind assertion. This is an
2195 extension compared with Perl, which requires all branches to match the same
2196 length of string. An assertion such as
2197 <pre>
2198 (?&#60;=ab(c|de))
2199 </pre>
2200 is not permitted, because its single top-level branch can match two different
2201 lengths, but it is acceptable to PCRE if rewritten to use two top-level
2202 branches:
2203 <pre>
2204 (?&#60;=abc|abde)
2205 </pre>
2206 In some cases, the escape sequence \K
2207 <a href="#resetmatchstart">(see above)</a>
2208 can be used instead of a lookbehind assertion to get round the fixed-length
2209 restriction.
2210 </P>
2211 <P>
2212 The implementation of lookbehind assertions is, for each alternative, to
2213 temporarily move the current position back by the fixed length and then try to
2214 match. If there are insufficient characters before the current position, the
2215 assertion fails.
2216 </P>
2217 <P>
2218 In a UTF mode, PCRE does not allow the \C escape (which matches a single data
2219 unit even in a UTF mode) to appear in lookbehind assertions, because it makes
2220 it impossible to calculate the length of the lookbehind. The \X and \R
2221 escapes, which can match different numbers of data units, are also not
2222 permitted.
2223 </P>
2224 <P>
2225 <a href="#subpatternsassubroutines">"Subroutine"</a>
2226 calls (see below) such as (?2) or (?&X) are permitted in lookbehinds, as long
2227 as the subpattern matches a fixed-length string.
2228 <a href="#recursion">Recursion,</a>
2229 however, is not supported.
2230 </P>
2231 <P>
2232 Possessive quantifiers can be used in conjunction with lookbehind assertions to
2233 specify efficient matching of fixed-length strings at the end of subject
2234 strings. Consider a simple pattern such as
2235 <pre>
2236 abcd$
2237 </pre>
2238 when applied to a long string that does not match. Because matching proceeds
2239 from left to right, PCRE will look for each "a" in the subject and then see if
2240 what follows matches the rest of the pattern. If the pattern is specified as
2241 <pre>
2242 ^.*abcd$
2243 </pre>
2244 the initial .* matches the entire string at first, but when this fails (because
2245 there is no following "a"), it backtracks to match all but the last character,
2246 then all but the last two characters, and so on. Once again the search for "a"
2247 covers the entire string, from right to left, so we are no better off. However,
2248 if the pattern is written as
2249 <pre>
2250 ^.*+(?&#60;=abcd)
2251 </pre>
2252 there can be no backtracking for the .*+ item; it can match only the entire
2253 string. The subsequent lookbehind assertion does a single test on the last four
2254 characters. If it fails, the match fails immediately. For long strings, this
2255 approach makes a significant difference to the processing time.
2256 </P>
2257 <br><b>
2258 Using multiple assertions
2259 </b><br>
2260 <P>
2261 Several assertions (of any sort) may occur in succession. For example,
2262 <pre>
2263 (?&#60;=\d{3})(?&#60;!999)foo
2264 </pre>
2265 matches "foo" preceded by three digits that are not "999". Notice that each of
2266 the assertions is applied independently at the same point in the subject
2267 string. First there is a check that the previous three characters are all
2268 digits, and then there is a check that the same three characters are not "999".
2269 This pattern does <i>not</i> match "foo" preceded by six characters, the first
2270 of which are digits and the last three of which are not "999". For example, it
2271 doesn't match "123abcfoo". A pattern to do that is
2272 <pre>
2273 (?&#60;=\d{3}...)(?&#60;!999)foo
2274 </pre>
2275 This time the first assertion looks at the preceding six characters, checking
2276 that the first three are digits, and then the second assertion checks that the
2277 preceding three characters are not "999".
2278 </P>
2279 <P>
2280 Assertions can be nested in any combination. For example,
2281 <pre>
2282 (?&#60;=(?&#60;!foo)bar)baz
2283 </pre>
2284 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
2285 preceded by "foo", while
2286 <pre>
2287 (?&#60;=\d{3}(?!999)...)foo
2288 </pre>
2289 is another pattern that matches "foo" preceded by three digits and any three
2290 characters that are not "999".
2291 <a name="conditions"></a></P>
2292 <br><a name="SEC21" href="#TOC1">CONDITIONAL SUBPATTERNS</a><br>
2293 <P>
2294 It is possible to cause the matching process to obey a subpattern
2295 conditionally or to choose between two alternative subpatterns, depending on
2296 the result of an assertion, or whether a specific capturing subpattern has
2297 already been matched. The two possible forms of conditional subpattern are:
2298 <pre>
2299 (?(condition)yes-pattern)
2300 (?(condition)yes-pattern|no-pattern)
2301 </pre>
2302 If the condition is satisfied, the yes-pattern is used; otherwise the
2303 no-pattern (if present) is used. If there are more than two alternatives in the
2304 subpattern, a compile-time error occurs. Each of the two alternatives may
2305 itself contain nested subpatterns of any form, including conditional
2306 subpatterns; the restriction to two alternatives applies only at the level of
2307 the condition. This pattern fragment is an example where the alternatives are
2308 complex:
2309 <pre>
2310 (?(1) (A|B|C) | (D | (?(2)E|F) | E) )
2312 </PRE>
2313 </P>
2314 <P>
2315 There are four kinds of condition: references to subpatterns, references to
2316 recursion, a pseudo-condition called DEFINE, and assertions.
2317 </P>
2318 <br><b>
2319 Checking for a used subpattern by number
2320 </b><br>
2321 <P>
2322 If the text between the parentheses consists of a sequence of digits, the
2323 condition is true if a capturing subpattern of that number has previously
2324 matched. If there is more than one capturing subpattern with the same number
2325 (see the earlier
2326 <a href="#recursion">section about duplicate subpattern numbers),</a>
2327 the condition is true if any of them have matched. An alternative notation is
2328 to precede the digits with a plus or minus sign. In this case, the subpattern
2329 number is relative rather than absolute. The most recently opened parentheses
2330 can be referenced by (?(-1), the next most recent by (?(-2), and so on. Inside
2331 loops it can also make sense to refer to subsequent groups. The next
2332 parentheses to be opened can be referenced as (?(+1), and so on. (The value
2333 zero in any of these forms is not used; it provokes a compile-time error.)
2334 </P>
2335 <P>
2336 Consider the following pattern, which contains non-significant white space to
2337 make it more readable (assume the PCRE_EXTENDED option) and to divide it into
2338 three parts for ease of discussion:
2339 <pre>
2340 ( \( )? [^()]+ (?(1) \) )
2341 </pre>
2342 The first part matches an optional opening parenthesis, and if that
2343 character is present, sets it as the first captured substring. The second part
2344 matches one or more characters that are not parentheses. The third part is a
2345 conditional subpattern that tests whether or not the first set of parentheses
2346 matched. If they did, that is, if subject started with an opening parenthesis,
2347 the condition is true, and so the yes-pattern is executed and a closing
2348 parenthesis is required. Otherwise, since no-pattern is not present, the
2349 subpattern matches nothing. In other words, this pattern matches a sequence of
2350 non-parentheses, optionally enclosed in parentheses.
2351 </P>
2352 <P>
2353 If you were embedding this pattern in a larger one, you could use a relative
2354 reference:
2355 <pre>
2356 ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
2357 </pre>
2358 This makes the fragment independent of the parentheses in the larger pattern.
2359 </P>
2360 <br><b>
2361 Checking for a used subpattern by name
2362 </b><br>
2363 <P>
2364 Perl uses the syntax (?(&#60;name&#62;)...) or (?('name')...) to test for a used
2365 subpattern by name. For compatibility with earlier versions of PCRE, which had
2366 this facility before Perl, the syntax (?(name)...) is also recognized.
2367 </P>
2368 <P>
2369 Rewriting the above example to use a named subpattern gives this:
2370 <pre>
2371 (?&#60;OPEN&#62; \( )? [^()]+ (?(&#60;OPEN&#62;) \) )
2372 </pre>
2373 If the name used in a condition of this kind is a duplicate, the test is
2374 applied to all subpatterns of the same name, and is true if any one of them has
2375 matched.
2376 </P>
2377 <br><b>
2378 Checking for pattern recursion
2379 </b><br>
2380 <P>
2381 If the condition is the string (R), and there is no subpattern with the name R,
2382 the condition is true if a recursive call to the whole pattern or any
2383 subpattern has been made. If digits or a name preceded by ampersand follow the
2384 letter R, for example:
2385 <pre>
2386 (?(R3)...) or (?(R&name)...)
2387 </pre>
2388 the condition is true if the most recent recursion is into a subpattern whose
2389 number or name is given. This condition does not check the entire recursion
2390 stack. If the name used in a condition of this kind is a duplicate, the test is
2391 applied to all subpatterns of the same name, and is true if any one of them is
2392 the most recent recursion.
2393 </P>
2394 <P>
2395 At "top level", all these recursion test conditions are false.
2396 <a href="#recursion">The syntax for recursive patterns</a>
2397 is described below.
2398 <a name="subdefine"></a></P>
2399 <br><b>
2400 Defining subpatterns for use by reference only
2401 </b><br>
2402 <P>
2403 If the condition is the string (DEFINE), and there is no subpattern with the
2404 name DEFINE, the condition is always false. In this case, there may be only one
2405 alternative in the subpattern. It is always skipped if control reaches this
2406 point in the pattern; the idea of DEFINE is that it can be used to define
2407 subroutines that can be referenced from elsewhere. (The use of
2408 <a href="#subpatternsassubroutines">subroutines</a>
2409 is described below.) For example, a pattern to match an IPv4 address such as
2410 "" could be written like this (ignore white space and line
2411 breaks):
2412 <pre>
2413 (?(DEFINE) (?&#60;byte&#62; 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
2414 \b (?&byte) (\.(?&byte)){3} \b
2415 </pre>
2416 The first part of the pattern is a DEFINE group inside which a another group
2417 named "byte" is defined. This matches an individual component of an IPv4
2418 address (a number less than 256). When matching takes place, this part of the
2419 pattern is skipped because DEFINE acts like a false condition. The rest of the
2420 pattern uses references to the named group to match the four dot-separated
2421 components of an IPv4 address, insisting on a word boundary at each end.
2422 </P>
2423 <br><b>
2424 Assertion conditions
2425 </b><br>
2426 <P>
2427 If the condition is not in any of the above formats, it must be an assertion.
2428 This may be a positive or negative lookahead or lookbehind assertion. Consider
2429 this pattern, again containing non-significant white space, and with the two
2430 alternatives on the second line:
2431 <pre>
2432 (?(?=[^a-z]*[a-z])
2433 \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
2434 </pre>
2435 The condition is a positive lookahead assertion that matches an optional
2436 sequence of non-letters followed by a letter. In other words, it tests for the
2437 presence of at least one letter in the subject. If a letter is found, the
2438 subject is matched against the first alternative; otherwise it is matched
2439 against the second. This pattern matches strings in one of the two forms
2440 dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
2441 <a name="comments"></a></P>
2442 <br><a name="SEC22" href="#TOC1">COMMENTS</a><br>
2443 <P>
2444 There are two ways of including comments in patterns that are processed by
2445 PCRE. In both cases, the start of the comment must not be in a character class,
2446 nor in the middle of any other sequence of related characters such as (?: or a
2447 subpattern name or number. The characters that make up a comment play no part
2448 in the pattern matching.
2449 </P>
2450 <P>
2451 The sequence (?# marks the start of a comment that continues up to the next
2452 closing parenthesis. Nested parentheses are not permitted. If the PCRE_EXTENDED
2453 option is set, an unescaped # character also introduces a comment, which in
2454 this case continues to immediately after the next newline character or
2455 character sequence in the pattern. Which characters are interpreted as newlines
2456 is controlled by the options passed to a compiling function or by a special
2457 sequence at the start of the pattern, as described in the section entitled
2458 <a href="#newlines">"Newline conventions"</a>
2459 above. Note that the end of this type of comment is a literal newline sequence
2460 in the pattern; escape sequences that happen to represent a newline do not
2461 count. For example, consider this pattern when PCRE_EXTENDED is set, and the
2462 default newline convention is in force:
2463 <pre>
2464 abc #comment \n still comment
2465 </pre>
2466 On encountering the # character, <b>pcre_compile()</b> skips along, looking for
2467 a newline in the pattern. The sequence \n is still literal at this stage, so
2468 it does not terminate the comment. Only an actual character with the code value
2469 0x0a (the default newline) does so.
2470 <a name="recursion"></a></P>
2471 <br><a name="SEC23" href="#TOC1">RECURSIVE PATTERNS</a><br>
2472 <P>
2473 Consider the problem of matching a string in parentheses, allowing for
2474 unlimited nested parentheses. Without the use of recursion, the best that can
2475 be done is to use a pattern that matches up to some fixed depth of nesting. It
2476 is not possible to handle an arbitrary nesting depth.
2477 </P>
2478 <P>
2479 For some time, Perl has provided a facility that allows regular expressions to
2480 recurse (amongst other things). It does this by interpolating Perl code in the
2481 expression at run time, and the code can refer to the expression itself. A Perl
2482 pattern using code interpolation to solve the parentheses problem can be
2483 created like this:
2484 <pre>
2485 $re = qr{\( (?: (?&#62;[^()]+) | (?p{$re}) )* \)}x;
2486 </pre>
2487 The (?p{...}) item interpolates Perl code at run time, and in this case refers
2488 recursively to the pattern in which it appears.
2489 </P>
2490 <P>
2491 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
2492 supports special syntax for recursion of the entire pattern, and also for
2493 individual subpattern recursion. After its introduction in PCRE and Python,
2494 this kind of recursion was subsequently introduced into Perl at release 5.10.
2495 </P>
2496 <P>
2497 A special item that consists of (? followed by a number greater than zero and a
2498 closing parenthesis is a recursive subroutine call of the subpattern of the
2499 given number, provided that it occurs inside that subpattern. (If not, it is a
2500 <a href="#subpatternsassubroutines">non-recursive subroutine</a>
2501 call, which is described in the next section.) The special item (?R) or (?0) is
2502 a recursive call of the entire regular expression.
2503 </P>
2504 <P>
2505 This PCRE pattern solves the nested parentheses problem (assume the
2506 PCRE_EXTENDED option is set so that white space is ignored):
2507 <pre>
2508 \( ( [^()]++ | (?R) )* \)
2509 </pre>
2510 First it matches an opening parenthesis. Then it matches any number of
2511 substrings which can either be a sequence of non-parentheses, or a recursive
2512 match of the pattern itself (that is, a correctly parenthesized substring).
2513 Finally there is a closing parenthesis. Note the use of a possessive quantifier
2514 to avoid backtracking into sequences of non-parentheses.
2515 </P>
2516 <P>
2517 If this were part of a larger pattern, you would not want to recurse the entire
2518 pattern, so instead you could use this:
2519 <pre>
2520 ( \( ( [^()]++ | (?1) )* \) )
2521 </pre>
2522 We have put the pattern into parentheses, and caused the recursion to refer to
2523 them instead of the whole pattern.
2524 </P>
2525 <P>
2526 In a larger pattern, keeping track of parenthesis numbers can be tricky. This
2527 is made easier by the use of relative references. Instead of (?1) in the
2528 pattern above you can write (?-2) to refer to the second most recently opened
2529 parentheses preceding the recursion. In other words, a negative number counts
2530 capturing parentheses leftwards from the point at which it is encountered.
2531 </P>
2532 <P>
2533 It is also possible to refer to subsequently opened parentheses, by writing
2534 references such as (?+2). However, these cannot be recursive because the
2535 reference is not inside the parentheses that are referenced. They are always
2536 <a href="#subpatternsassubroutines">non-recursive subroutine</a>
2537 calls, as described in the next section.
2538 </P>
2539 <P>
2540 An alternative approach is to use named parentheses instead. The Perl syntax
2541 for this is (?&name); PCRE's earlier syntax (?P&#62;name) is also supported. We
2542 could rewrite the above example as follows:
2543 <pre>
2544 (?&#60;pn&#62; \( ( [^()]++ | (?&pn) )* \) )
2545 </pre>
2546 If there is more than one subpattern with the same name, the earliest one is
2547 used.
2548 </P>
2549 <P>
2550 This particular example pattern that we have been looking at contains nested
2551 unlimited repeats, and so the use of a possessive quantifier for matching
2552 strings of non-parentheses is important when applying the pattern to strings
2553 that do not match. For example, when this pattern is applied to
2554 <pre>
2555 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
2556 </pre>
2557 it yields "no match" quickly. However, if a possessive quantifier is not used,
2558 the match runs for a very long time indeed because there are so many different
2559 ways the + and * repeats can carve up the subject, and all have to be tested
2560 before failure can be reported.
2561 </P>
2562 <P>
2563 At the end of a match, the values of capturing parentheses are those from
2564 the outermost level. If you want to obtain intermediate values, a callout
2565 function can be used (see below and the
2566 <a href="pcrecallout.html"><b>pcrecallout</b></a>
2567 documentation). If the pattern above is matched against
2568 <pre>
2569 (ab(cd)ef)
2570 </pre>
2571 the value for the inner capturing parentheses (numbered 2) is "ef", which is
2572 the last value taken on at the top level. If a capturing subpattern is not
2573 matched at the top level, its final captured value is unset, even if it was
2574 (temporarily) set at a deeper level during the matching process.
2575 </P>
2576 <P>
2577 If there are more than 15 capturing parentheses in a pattern, PCRE has to
2578 obtain extra memory to store data during a recursion, which it does by using
2579 <b>pcre_malloc</b>, freeing it via <b>pcre_free</b> afterwards. If no memory can
2580 be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
2581 </P>
2582 <P>
2583 Do not confuse the (?R) item with the condition (R), which tests for recursion.
2584 Consider this pattern, which matches text in angle brackets, allowing for
2585 arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2586 recursing), whereas any characters are permitted at the outer level.
2587 <pre>
2588 &#60; (?: (?(R) \d++ | [^&#60;&#62;]*+) | (?R)) * &#62;
2589 </pre>
2590 In this pattern, (?(R) is the start of a conditional subpattern, with two
2591 different alternatives for the recursive and non-recursive cases. The (?R) item
2592 is the actual recursive call.
2593 <a name="recursiondifference"></a></P>
2594 <br><b>
2595 Differences in recursion processing between PCRE and Perl
2596 </b><br>
2597 <P>
2598 Recursion processing in PCRE differs from Perl in two important ways. In PCRE
2599 (like Python, but unlike Perl), a recursive subpattern call is always treated
2600 as an atomic group. That is, once it has matched some of the subject string, it
2601 is never re-entered, even if it contains untried alternatives and there is a
2602 subsequent matching failure. This can be illustrated by the following pattern,
2603 which purports to match a palindromic string that contains an odd number of
2604 characters (for example, "a", "aba", "abcba", "abcdcba"):
2605 <pre>
2606 ^(.|(.)(?1)\2)$
2607 </pre>
2608 The idea is that it either matches a single character, or two identical
2609 characters surrounding a sub-palindrome. In Perl, this pattern works; in PCRE
2610 it does not if the pattern is longer than three characters. Consider the
2611 subject string "abcba":
2612 </P>
2613 <P>
2614 At the top level, the first character is matched, but as it is not at the end
2615 of the string, the first alternative fails; the second alternative is taken
2616 and the recursion kicks in. The recursive call to subpattern 1 successfully
2617 matches the next character ("b"). (Note that the beginning and end of line
2618 tests are not part of the recursion).
2619 </P>
2620 <P>
2621 Back at the top level, the next character ("c") is compared with what
2622 subpattern 2 matched, which was "a". This fails. Because the recursion is
2623 treated as an atomic group, there are now no backtracking points, and so the
2624 entire match fails. (Perl is able, at this point, to re-enter the recursion and
2625 try the second alternative.) However, if the pattern is written with the
2626 alternatives in the other order, things are different:
2627 <pre>
2628 ^((.)(?1)\2|.)$
2629 </pre>
2630 This time, the recursing alternative is tried first, and continues to recurse
2631 until it runs out of characters, at which point the recursion fails. But this
2632 time we do have another alternative to try at the higher level. That is the big
2633 difference: in the previous case the remaining alternative is at a deeper
2634 recursion level, which PCRE cannot use.
2635 </P>
2636 <P>
2637 To change the pattern so that it matches all palindromic strings, not just
2638 those with an odd number of characters, it is tempting to change the pattern to
2639 this:
2640 <pre>
2641 ^((.)(?1)\2|.?)$
2642 </pre>
2643 Again, this works in Perl, but not in PCRE, and for the same reason. When a
2644 deeper recursion has matched a single character, it cannot be entered again in
2645 order to match an empty string. The solution is to separate the two cases, and
2646 write out the odd and even cases as alternatives at the higher level:
2647 <pre>
2648 ^(?:((.)(?1)\2|)|((.)(?3)\4|.))
2649 </pre>
2650 If you want to match typical palindromic phrases, the pattern has to ignore all
2651 non-word characters, which can be done like this:
2652 <pre>
2653 ^\W*+(?:((.)\W*+(?1)\W*+\2|)|((.)\W*+(?3)\W*+\4|\W*+.\W*+))\W*+$
2654 </pre>
2655 If run with the PCRE_CASELESS option, this pattern matches phrases such as "A
2656 man, a plan, a canal: Panama!" and it works well in both PCRE and Perl. Note
2657 the use of the possessive quantifier *+ to avoid backtracking into sequences of
2658 non-word characters. Without this, PCRE takes a great deal longer (ten times or
2659 more) to match typical phrases, and Perl takes so long that you think it has
2660 gone into a loop.
2661 </P>
2662 <P>
2663 <b>WARNING</b>: The palindrome-matching patterns above work only if the subject
2664 string does not start with a palindrome that is shorter than the entire string.
2665 For example, although "abcba" is correctly matched, if the subject is "ababa",
2666 PCRE finds the palindrome "aba" at the start, then fails at top level because
2667 the end of the string does not follow. Once again, it cannot jump back into the
2668 recursion to try other alternatives, so the entire match fails.
2669 </P>
2670 <P>
2671 The second way in which PCRE and Perl differ in their recursion processing is
2672 in the handling of captured values. In Perl, when a subpattern is called
2673 recursively or as a subpattern (see the next section), it has no access to any
2674 values that were captured outside the recursion, whereas in PCRE these values
2675 can be referenced. Consider this pattern:
2676 <pre>
2677 ^(.)(\1|a(?2))
2678 </pre>
2679 In PCRE, this pattern matches "bab". The first capturing parentheses match "b",
2680 then in the second group, when the back reference \1 fails to match "b", the
2681 second alternative matches "a" and then recurses. In the recursion, \1 does
2682 now match "b" and so the whole match succeeds. In Perl, the pattern fails to
2683 match because inside the recursive call \1 cannot access the externally set
2684 value.
2685 <a name="subpatternsassubroutines"></a></P>
2686 <br><a name="SEC24" href="#TOC1">SUBPATTERNS AS SUBROUTINES</a><br>
2687 <P>
2688 If the syntax for a recursive subpattern call (either by number or by
2689 name) is used outside the parentheses to which it refers, it operates like a
2690 subroutine in a programming language. The called subpattern may be defined
2691 before or after the reference. A numbered reference can be absolute or
2692 relative, as in these examples:
2693 <pre>
2694 (...(absolute)...)...(?2)...
2695 (...(relative)...)...(?-1)...
2696 (...(?+1)...(relative)...
2697 </pre>
2698 An earlier example pointed out that the pattern
2699 <pre>
2700 (sens|respons)e and \1ibility
2701 </pre>
2702 matches "sense and sensibility" and "response and responsibility", but not
2703 "sense and responsibility". If instead the pattern
2704 <pre>
2705 (sens|respons)e and (?1)ibility
2706 </pre>
2707 is used, it does match "sense and responsibility" as well as the other two
2708 strings. Another example is given in the discussion of DEFINE above.
2709 </P>
2710 <P>
2711 All subroutine calls, whether recursive or not, are always treated as atomic
2712 groups. That is, once a subroutine has matched some of the subject string, it
2713 is never re-entered, even if it contains untried alternatives and there is a
2714 subsequent matching failure. Any capturing parentheses that are set during the
2715 subroutine call revert to their previous values afterwards.
2716 </P>
2717 <P>
2718 Processing options such as case-independence are fixed when a subpattern is
2719 defined, so if it is used as a subroutine, such options cannot be changed for
2720 different calls. For example, consider this pattern:
2721 <pre>
2722 (abc)(?i:(?-1))
2723 </pre>
2724 It matches "abcabc". It does not match "abcABC" because the change of
2725 processing option does not affect the called subpattern.
2726 <a name="onigurumasubroutines"></a></P>
2727 <br><a name="SEC25" href="#TOC1">ONIGURUMA SUBROUTINE SYNTAX</a><br>
2728 <P>
2729 For compatibility with Oniguruma, the non-Perl syntax \g followed by a name or
2730 a number enclosed either in angle brackets or single quotes, is an alternative
2731 syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2732 are two of the examples used above, rewritten using this syntax:
2733 <pre>
2734 (?&#60;pn&#62; \( ( (?&#62;[^()]+) | \g&#60;pn&#62; )* \) )
2735 (sens|respons)e and \g'1'ibility
2736 </pre>
2737 PCRE supports an extension to Oniguruma: if a number is preceded by a
2738 plus or a minus sign it is taken as a relative reference. For example:
2739 <pre>
2740 (abc)(?i:\g&#60;-1&#62;)
2741 </pre>
2742 Note that \g{...} (Perl syntax) and \g&#60;...&#62; (Oniguruma syntax) are <i>not</i>
2743 synonymous. The former is a back reference; the latter is a subroutine call.
2744 </P>
2745 <br><a name="SEC26" href="#TOC1">CALLOUTS</a><br>
2746 <P>
2747 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2748 code to be obeyed in the middle of matching a regular expression. This makes it
2749 possible, amongst other things, to extract different substrings that match the
2750 same pair of parentheses when there is a repetition.
2751 </P>
2752 <P>
2753 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2754 code. The feature is called "callout". The caller of PCRE provides an external
2755 function by putting its entry point in the global variable <i>pcre_callout</i>
2756 (8-bit library) or <i>pcre[16|32]_callout</i> (16-bit or 32-bit library).
2757 By default, this variable contains NULL, which disables all calling out.
2758 </P>
2759 <P>
2760 Within a regular expression, (?C) indicates the points at which the external
2761 function is to be called. If you want to identify different callout points, you
2762 can put a number less than 256 after the letter C. The default value is zero.
2763 For example, this pattern has two callout points:
2764 <pre>
2765 (?C1)abc(?C2)def
2766 </pre>
2767 If the PCRE_AUTO_CALLOUT flag is passed to a compiling function, callouts are
2768 automatically installed before each item in the pattern. They are all numbered
2769 255. If there is a conditional group in the pattern whose condition is an
2770 assertion, an additional callout is inserted just before the condition. An
2771 explicit callout may also be set at this position, as in this example:
2772 <pre>
2773 (?(?C9)(?=a)abc|def)
2774 </pre>
2775 Note that this applies only to assertion conditions, not to other types of
2776 condition.
2777 </P>
2778 <P>
2779 During matching, when PCRE reaches a callout point, the external function is
2780 called. It is provided with the number of the callout, the position in the
2781 pattern, and, optionally, one item of data originally supplied by the caller of
2782 the matching function. The callout function may cause matching to proceed, to
2783 backtrack, or to fail altogether.
2784 </P>
2785 <P>
2786 By default, PCRE implements a number of optimizations at compile time and
2787 matching time, and one side-effect is that sometimes callouts are skipped. If
2788 you need all possible callouts to happen, you need to set options that disable
2789 the relevant optimizations. More details, and a complete description of the
2790 interface to the callout function, are given in the
2791 <a href="pcrecallout.html"><b>pcrecallout</b></a>
2792 documentation.
2793 <a name="backtrackcontrol"></a></P>
2794 <br><a name="SEC27" href="#TOC1">BACKTRACKING CONTROL</a><br>
2795 <P>
2796 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2797 are still described in the Perl documentation as "experimental and subject to
2798 change or removal in a future version of Perl". It goes on to say: "Their usage
2799 in production code should be noted to avoid problems during upgrades." The same
2800 remarks apply to the PCRE features described in this section.
2801 </P>
2802 <P>
2803 The new verbs make use of what was previously invalid syntax: an opening
2804 parenthesis followed by an asterisk. They are generally of the form
2805 (*VERB) or (*VERB:NAME). Some may take either form, possibly behaving
2806 differently depending on whether or not a name is present. A name is any
2807 sequence of characters that does not include a closing parenthesis. The maximum
2808 length of name is 255 in the 8-bit library and 65535 in the 16-bit and 32-bit
2809 libraries. If the name is empty, that is, if the closing parenthesis
2810 immediately follows the colon, the effect is as if the colon were not there.
2811 Any number of these verbs may occur in a pattern.
2812 </P>
2813 <P>
2814 Since these verbs are specifically related to backtracking, most of them can be
2815 used only when the pattern is to be matched using one of the traditional
2816 matching functions, because these use a backtracking algorithm. With the
2817 exception of (*FAIL), which behaves like a failing negative assertion, the
2818 backtracking control verbs cause an error if encountered by a DFA matching
2819 function.
2820 </P>
2821 <P>
2822 The behaviour of these verbs in
2823 <a href="#btrepeat">repeated groups,</a>
2824 <a href="#btassert">assertions,</a>
2825 and in
2826 <a href="#btsub">subpatterns called as subroutines</a>
2827 (whether or not recursively) is documented below.
2828 <a name="nooptimize"></a></P>
2829 <br><b>
2830 Optimizations that affect backtracking verbs
2831 </b><br>
2832 <P>
2833 PCRE contains some optimizations that are used to speed up matching by running
2834 some checks at the start of each match attempt. For example, it may know the
2835 minimum length of matching subject, or that a particular character must be
2836 present. When one of these optimizations bypasses the running of a match, any
2837 included backtracking verbs will not, of course, be processed. You can suppress
2838 the start-of-match optimizations by setting the PCRE_NO_START_OPTIMIZE option
2839 when calling <b>pcre_compile()</b> or <b>pcre_exec()</b>, or by starting the
2840 pattern with (*NO_START_OPT). There is more discussion of this option in the
2841 section entitled
2842 <a href="pcreapi.html#execoptions">"Option bits for <b>pcre_exec()</b>"</a>
2843 in the
2844 <a href="pcreapi.html"><b>pcreapi</b></a>
2845 documentation.
2846 </P>
2847 <P>
2848 Experiments with Perl suggest that it too has similar optimizations, sometimes
2849 leading to anomalous results.
2850 </P>
2851 <br><b>
2852 Verbs that act immediately
2853 </b><br>
2854 <P>
2855 The following verbs act as soon as they are encountered. They may not be
2856 followed by a name.
2857 <pre>
2858 (*ACCEPT)
2859 </pre>
2860 This verb causes the match to end successfully, skipping the remainder of the
2861 pattern. However, when it is inside a subpattern that is called as a
2862 subroutine, only that subpattern is ended successfully. Matching then continues
2863 at the outer level. If (*ACCEPT) in triggered in a positive assertion, the
2864 assertion succeeds; in a negative assertion, the assertion fails.
2865 </P>
2866 <P>
2867 If (*ACCEPT) is inside capturing parentheses, the data so far is captured. For
2868 example:
2869 <pre>
2870 A((?:A|B(*ACCEPT)|C)D)
2871 </pre>
2872 This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is captured by
2873 the outer parentheses.
2874 <pre>
2875 (*FAIL) or (*F)
2876 </pre>
2877 This verb causes a matching failure, forcing backtracking to occur. It is
2878 equivalent to (?!) but easier to read. The Perl documentation notes that it is
2879 probably useful only when combined with (?{}) or (??{}). Those are, of course,
2880 Perl features that are not present in PCRE. The nearest equivalent is the
2881 callout feature, as for example in this pattern:
2882 <pre>
2883 a+(?C)(*FAIL)
2884 </pre>
2885 A match with the string "aaaa" always fails, but the callout is taken before
2886 each backtrack happens (in this example, 10 times).
2887 </P>
2888 <br><b>
2889 Recording which path was taken
2890 </b><br>
2891 <P>
2892 There is one verb whose main purpose is to track how a match was arrived at,
2893 though it also has a secondary use in conjunction with advancing the match
2894 starting point (see (*SKIP) below).
2895 <pre>
2896 (*MARK:NAME) or (*:NAME)
2897 </pre>
2898 A name is always required with this verb. There may be as many instances of
2899 (*MARK) as you like in a pattern, and their names do not have to be unique.
2900 </P>
2901 <P>
2902 When a match succeeds, the name of the last-encountered (*MARK:NAME),
2903 (*PRUNE:NAME), or (*THEN:NAME) on the matching path is passed back to the
2904 caller as described in the section entitled
2905 <a href="pcreapi.html#extradata">"Extra data for <b>pcre_exec()</b>"</a>
2906 in the
2907 <a href="pcreapi.html"><b>pcreapi</b></a>
2908 documentation. Here is an example of <b>pcretest</b> output, where the /K
2909 modifier requests the retrieval and outputting of (*MARK) data:
2910 <pre>
2911 re&#62; /X(*MARK:A)Y|X(*MARK:B)Z/K
2912 data&#62; XY
2913 0: XY
2914 MK: A
2915 XZ
2916 0: XZ
2917 MK: B
2918 </pre>
2919 The (*MARK) name is tagged with "MK:" in this output, and in this example it
2920 indicates which of the two alternatives matched. This is a more efficient way
2921 of obtaining this information than putting each alternative in its own
2922 capturing parentheses.
2923 </P>
2924 <P>
2925 If a verb with a name is encountered in a positive assertion that is true, the
2926 name is recorded and passed back if it is the last-encountered. This does not
2927 happen for negative assertions or failing positive assertions.
2928 </P>
2929 <P>
2930 After a partial match or a failed match, the last encountered name in the
2931 entire match process is returned. For example:
2932 <pre>
2933 re&#62; /X(*MARK:A)Y|X(*MARK:B)Z/K
2934 data&#62; XP
2935 No match, mark = B
2936 </pre>
2937 Note that in this unanchored example the mark is retained from the match
2938 attempt that started at the letter "X" in the subject. Subsequent match
2939 attempts starting at "P" and then with an empty string do not get as far as the
2940 (*MARK) item, but nevertheless do not reset it.
2941 </P>
2942 <P>
2943 If you are interested in (*MARK) values after failed matches, you should
2944 probably set the PCRE_NO_START_OPTIMIZE option
2945 <a href="#nooptimize">(see above)</a>
2946 to ensure that the match is always attempted.
2947 </P>
2948 <br><b>
2949 Verbs that act after backtracking
2950 </b><br>
2951 <P>
2952 The following verbs do nothing when they are encountered. Matching continues
2953 with what follows, but if there is no subsequent match, causing a backtrack to
2954 the verb, a failure is forced. That is, backtracking cannot pass to the left of
2955 the verb. However, when one of these verbs appears inside an atomic group or an
2956 assertion that is true, its effect is confined to that group, because once the
2957 group has been matched, there is never any backtracking into it. In this
2958 situation, backtracking can "jump back" to the left of the entire atomic group
2959 or assertion. (Remember also, as stated above, that this localization also
2960 applies in subroutine calls.)
2961 </P>
2962 <P>
2963 These verbs differ in exactly what kind of failure occurs when backtracking
2964 reaches them. The behaviour described below is what happens when the verb is
2965 not in a subroutine or an assertion. Subsequent sections cover these special
2966 cases.
2967 <pre>
2968 (*COMMIT)
2969 </pre>
2970 This verb, which may not be followed by a name, causes the whole match to fail
2971 outright if there is a later matching failure that causes backtracking to reach
2972 it. Even if the pattern is unanchored, no further attempts to find a match by
2973 advancing the starting point take place. If (*COMMIT) is the only backtracking
2974 verb that is encountered, once it has been passed <b>pcre_exec()</b> is
2975 committed to finding a match at the current starting point, or not at all. For
2976 example:
2977 <pre>
2978 a+(*COMMIT)b
2979 </pre>
2980 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2981 dynamic anchor, or "I've started, so I must finish." The name of the most
2982 recently passed (*MARK) in the path is passed back when (*COMMIT) forces a
2983 match failure.
2984 </P>
2985 <P>
2986 If there is more than one backtracking verb in a pattern, a different one that
2987 follows (*COMMIT) may be triggered first, so merely passing (*COMMIT) during a
2988 match does not always guarantee that a match must be at this starting point.
2989 </P>
2990 <P>
2991 Note that (*COMMIT) at the start of a pattern is not the same as an anchor,
2992 unless PCRE's start-of-match optimizations are turned off, as shown in this
2993 <b>pcretest</b> example:
2994 <pre>
2995 re&#62; /(*COMMIT)abc/
2996 data&#62; xyzabc
2997 0: abc
2998 xyzabc\Y
2999 No match
3000 </pre>
3001 PCRE knows that any match must start with "a", so the optimization skips along
3002 the subject to "a" before running the first match attempt, which succeeds. When
3003 the optimization is disabled by the \Y escape in the second subject, the match
3004 starts at "x" and so the (*COMMIT) causes it to fail without trying any other
3005 starting points.
3006 <pre>
3007 (*PRUNE) or (*PRUNE:NAME)
3008 </pre>
3009 This verb causes the match to fail at the current starting position in the
3010 subject if there is a later matching failure that causes backtracking to reach
3011 it. If the pattern is unanchored, the normal "bumpalong" advance to the next
3012 starting character then happens. Backtracking can occur as usual to the left of
3013 (*PRUNE), before it is reached, or when matching to the right of (*PRUNE), but
3014 if there is no match to the right, backtracking cannot cross (*PRUNE). In
3015 simple cases, the use of (*PRUNE) is just an alternative to an atomic group or
3016 possessive quantifier, but there are some uses of (*PRUNE) that cannot be
3017 expressed in any other way. In an anchored pattern (*PRUNE) has the same effect
3018 as (*COMMIT).
3019 </P>
3020 <P>
3021 The behaviour of (*PRUNE:NAME) is the not the same as (*MARK:NAME)(*PRUNE).
3022 It is like (*MARK:NAME) in that the name is remembered for passing back to the
3023 caller. However, (*SKIP:NAME) searches only for names set with (*MARK).
3024 <pre>
3025 (*SKIP)
3026 </pre>
3027 This verb, when given without a name, is like (*PRUNE), except that if the
3028 pattern is unanchored, the "bumpalong" advance is not to the next character,
3029 but to the position in the subject where (*SKIP) was encountered. (*SKIP)
3030 signifies that whatever text was matched leading up to it cannot be part of a
3031 successful match. Consider:
3032 <pre>
3033 a+(*SKIP)b
3034 </pre>
3035 If the subject is "aaaac...", after the first match attempt fails (starting at
3036 the first character in the string), the starting point skips on to start the
3037 next attempt at "c". Note that a possessive quantifer does not have the same
3038 effect as this example; although it would suppress backtracking during the
3039 first match attempt, the second attempt would start at the second character
3040 instead of skipping on to "c".
3041 <pre>
3042 (*SKIP:NAME)
3043 </pre>
3044 When (*SKIP) has an associated name, its behaviour is modified. When it is
3045 triggered, the previous path through the pattern is searched for the most
3046 recent (*MARK) that has the same name. If one is found, the "bumpalong" advance
3047 is to the subject position that corresponds to that (*MARK) instead of to where
3048 (*SKIP) was encountered. If no (*MARK) with a matching name is found, the
3049 (*SKIP) is ignored.
3050 </P>
3051 <P>
3052 Note that (*SKIP:NAME) searches only for names set by (*MARK:NAME). It ignores
3053 names that are set by (*PRUNE:NAME) or (*THEN:NAME).
3054 <pre>
3055 (*THEN) or (*THEN:NAME)
3056 </pre>
3057 This verb causes a skip to the next innermost alternative when backtracking
3058 reaches it. That is, it cancels any further backtracking within the current
3059 alternative. Its name comes from the observation that it can be used for a
3060 pattern-based if-then-else block:
3061 <pre>
3062 ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
3063 </pre>
3064 If the COND1 pattern matches, FOO is tried (and possibly further items after
3065 the end of the group if FOO succeeds); on failure, the matcher skips to the
3066 second alternative and tries COND2, without backtracking into COND1. If that
3067 succeeds and BAR fails, COND3 is tried. If subsequently BAZ fails, there are no
3068 more alternatives, so there is a backtrack to whatever came before the entire
3069 group. If (*THEN) is not inside an alternation, it acts like (*PRUNE).
3070 </P>
3071 <P>
3072 The behaviour of (*THEN:NAME) is the not the same as (*MARK:NAME)(*THEN).
3073 It is like (*MARK:NAME) in that the name is remembered for passing back to the
3074 caller. However, (*SKIP:NAME) searches only for names set with (*MARK).
3075 </P>
3076 <P>
3077 A subpattern that does not contain a | character is just a part of the
3078 enclosing alternative; it is not a nested alternation with only one
3079 alternative. The effect of (*THEN) extends beyond such a subpattern to the
3080 enclosing alternative. Consider this pattern, where A, B, etc. are complex
3081 pattern fragments that do not contain any | characters at this level:
3082 <pre>
3083 A (B(*THEN)C) | D
3084 </pre>
3085 If A and B are matched, but there is a failure in C, matching does not
3086 backtrack into A; instead it moves to the next alternative, that is, D.
3087 However, if the subpattern containing (*THEN) is given an alternative, it
3088 behaves differently:
3089 <pre>
3090 A (B(*THEN)C | (*FAIL)) | D
3091 </pre>
3092 The effect of (*THEN) is now confined to the inner subpattern. After a failure
3093 in C, matching moves to (*FAIL), which causes the whole subpattern to fail
3094 because there are no more alternatives to try. In this case, matching does now
3095 backtrack into A.
3096 </P>
3097 <P>
3098 Note that a conditional subpattern is not considered as having two
3099 alternatives, because only one is ever used. In other words, the | character in
3100 a conditional subpattern has a different meaning. Ignoring white space,
3101 consider:
3102 <pre>
3103 ^.*? (?(?=a) a | b(*THEN)c )
3104 </pre>
3105 If the subject is "ba", this pattern does not match. Because .*? is ungreedy,
3106 it initially matches zero characters. The condition (?=a) then fails, the
3107 character "b" is matched, but "c" is not. At this point, matching does not
3108 backtrack to .*? as might perhaps be expected from the presence of the |
3109 character. The conditional subpattern is part of the single alternative that
3110 comprises the whole pattern, and so the match fails. (If there was a backtrack
3111 into .*?, allowing it to match "b", the match would succeed.)
3112 </P>
3113 <P>
3114 The verbs just described provide four different "strengths" of control when
3115 subsequent matching fails. (*THEN) is the weakest, carrying on the match at the
3116 next alternative. (*PRUNE) comes next, failing the match at the current
3117 starting position, but allowing an advance to the next character (for an
3118 unanchored pattern). (*SKIP) is similar, except that the advance may be more
3119 than one character. (*COMMIT) is the strongest, causing the entire match to
3120 fail.
3121 </P>
3122 <br><b>
3123 More than one backtracking verb
3124 </b><br>
3125 <P>
3126 If more than one backtracking verb is present in a pattern, the one that is
3127 backtracked onto first acts. For example, consider this pattern, where A, B,
3128 etc. are complex pattern fragments:
3129 <pre>
3131 </pre>
3132 If A matches but B fails, the backtrack to (*COMMIT) causes the entire match to
3133 fail. However, if A and B match, but C fails, the backtrack to (*THEN) causes
3134 the next alternative (ABD) to be tried. This behaviour is consistent, but is
3135 not always the same as Perl's. It means that if two or more backtracking verbs
3136 appear in succession, all the the last of them has no effect. Consider this
3137 example:
3138 <pre>
3139 ...(*COMMIT)(*PRUNE)...
3140 </pre>
3141 If there is a matching failure to the right, backtracking onto (*PRUNE) causes
3142 it to be triggered, and its action is taken. There can never be a backtrack
3143 onto (*COMMIT).
3144 <a name="btrepeat"></a></P>
3145 <br><b>
3146 Backtracking verbs in repeated groups
3147 </b><br>
3148 <P>
3149 PCRE differs from Perl in its handling of backtracking verbs in repeated
3150 groups. For example, consider:
3151 <pre>
3152 /(a(*COMMIT)b)+ac/
3153 </pre>
3154 If the subject is "abac", Perl matches, but PCRE fails because the (*COMMIT) in
3155 the second repeat of the group acts.
3156 <a name="btassert"></a></P>
3157 <br><b>
3158 Backtracking verbs in assertions
3159 </b><br>
3160 <P>
3161 (*FAIL) in an assertion has its normal effect: it forces an immediate backtrack.
3162 </P>
3163 <P>
3164 (*ACCEPT) in a positive assertion causes the assertion to succeed without any
3165 further processing. In a negative assertion, (*ACCEPT) causes the assertion to
3166 fail without any further processing.
3167 </P>
3168 <P>
3169 The other backtracking verbs are not treated specially if they appear in a
3170 positive assertion. In particular, (*THEN) skips to the next alternative in the
3171 innermost enclosing group that has alternations, whether or not this is within
3172 the assertion.
3173 </P>
3174 <P>
3175 Negative assertions are, however, different, in order to ensure that changing a
3176 positive assertion into a negative assertion changes its result. Backtracking
3177 into (*COMMIT), (*SKIP), or (*PRUNE) causes a negative assertion to be true,
3178 without considering any further alternative branches in the assertion.
3179 Backtracking into (*THEN) causes it to skip to the next enclosing alternative
3180 within the assertion (the normal behaviour), but if the assertion does not have
3181 such an alternative, (*THEN) behaves like (*PRUNE).
3182 <a name="btsub"></a></P>
3183 <br><b>
3184 Backtracking verbs in subroutines
3185 </b><br>
3186 <P>
3187 These behaviours occur whether or not the subpattern is called recursively.
3188 Perl's treatment of subroutines is different in some cases.
3189 </P>
3190 <P>
3191 (*FAIL) in a subpattern called as a subroutine has its normal effect: it forces
3192 an immediate backtrack.
3193 </P>
3194 <P>
3195 (*ACCEPT) in a subpattern called as a subroutine causes the subroutine match to
3196 succeed without any further processing. Matching then continues after the
3197 subroutine call.
3198 </P>
3199 <P>
3200 (*COMMIT), (*SKIP), and (*PRUNE) in a subpattern called as a subroutine cause
3201 the subroutine match to fail.
3202 </P>
3203 <P>
3204 (*THEN) skips to the next alternative in the innermost enclosing group within
3205 the subpattern that has alternatives. If there is no such group within the
3206 subpattern, (*THEN) causes the subroutine match to fail.
3207 </P>
3208 <br><a name="SEC28" href="#TOC1">SEE ALSO</a><br>
3209 <P>
3210 <b>pcreapi</b>(3), <b>pcrecallout</b>(3), <b>pcrematching</b>(3),
3211 <b>pcresyntax</b>(3), <b>pcre</b>(3), <b>pcre16(3)</b>, <b>pcre32(3)</b>.
3212 </P>
3213 <br><a name="SEC29" href="#TOC1">AUTHOR</a><br>
3214 <P>
3215 Philip Hazel
3216 <br>
3217 University Computing Service
3218 <br>
3219 Cambridge CB2 3QH, England.
3220 <br>
3221 </P>
3222 <br><a name="SEC30" href="#TOC1">REVISION</a><br>
3223 <P>
3224 Last updated: 03 December 2013
3225 <br>
3226 Copyright &copy; 1997-2013 University of Cambridge.
3227 <br>
3228 <p>
3229 Return to the <a href="index.html">PCRE index page</a>.
3230 </p>


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

  ViewVC Help
Powered by ViewVC 1.1.5