Annotation of embedaddon/pcre/doc/pcrepattern.3, revision 1.1.1.1
1.1 misho 1: .TH PCREPATTERN 3
2: .SH NAME
3: PCRE - Perl-compatible regular expressions
4: .SH "PCRE REGULAR EXPRESSION DETAILS"
5: .rs
6: .sp
7: The syntax and semantics of the regular expressions that are supported by PCRE
8: are described in detail below. There is a quick-reference syntax summary in the
9: .\" HREF
10: \fBpcresyntax\fP
11: .\"
12: page. PCRE tries to match Perl syntax and semantics as closely as it can. PCRE
13: also supports some alternative regular expression syntax (which does not
14: conflict with the Perl syntax) in order to provide some compatibility with
15: regular expressions in Python, .NET, and Oniguruma.
16: .P
17: Perl's regular expressions are described in its own documentation, and
18: regular expressions in general are covered in a number of books, some of which
19: have copious examples. Jeffrey Friedl's "Mastering Regular Expressions",
20: published by O'Reilly, covers regular expressions in great detail. This
21: description of PCRE's regular expressions is intended as reference material.
22: .P
23: The original operation of PCRE was on strings of one-byte characters. However,
24: there is now also support for UTF-8 character strings. To use this,
25: PCRE must be built to include UTF-8 support, and you must call
26: \fBpcre_compile()\fP or \fBpcre_compile2()\fP with the PCRE_UTF8 option. There
27: is also a special sequence that can be given at the start of a pattern:
28: .sp
29: (*UTF8)
30: .sp
31: Starting a pattern with this sequence is equivalent to setting the PCRE_UTF8
32: option. This feature is not Perl-compatible. How setting UTF-8 mode affects
33: pattern matching is mentioned in several places below. There is also a summary
34: of UTF-8 features in the
35: .\" HREF
36: \fBpcreunicode\fP
37: .\"
38: page.
39: .P
40: Another special sequence that may appear at the start of a pattern or in
41: combination with (*UTF8) is:
42: .sp
43: (*UCP)
44: .sp
45: This has the same effect as setting the PCRE_UCP option: it causes sequences
46: such as \ed and \ew to use Unicode properties to determine character types,
47: instead of recognizing only characters with codes less than 128 via a lookup
48: table.
49: .P
50: If a pattern starts with (*NO_START_OPT), it has the same effect as setting the
51: PCRE_NO_START_OPTIMIZE option either at compile or matching time. There are
52: also some more of these special sequences that are concerned with the handling
53: of newlines; they are described below.
54: .P
55: The remainder of this document discusses the patterns that are supported by
56: PCRE when its main matching function, \fBpcre_exec()\fP, is used.
57: From release 6.0, PCRE offers a second matching function,
58: \fBpcre_dfa_exec()\fP, which matches using a different algorithm that is not
59: Perl-compatible. Some of the features discussed below are not available when
60: \fBpcre_dfa_exec()\fP is used. The advantages and disadvantages of the
61: alternative function, and how it differs from the normal function, are
62: discussed in the
63: .\" HREF
64: \fBpcrematching\fP
65: .\"
66: page.
67: .
68: .
69: .\" HTML <a name="newlines"></a>
70: .SH "NEWLINE CONVENTIONS"
71: .rs
72: .sp
73: PCRE supports five different conventions for indicating line breaks in
74: strings: a single CR (carriage return) character, a single LF (linefeed)
75: character, the two-character sequence CRLF, any of the three preceding, or any
76: Unicode newline sequence. The
77: .\" HREF
78: \fBpcreapi\fP
79: .\"
80: page has
81: .\" HTML <a href="pcreapi.html#newlines">
82: .\" </a>
83: further discussion
84: .\"
85: about newlines, and shows how to set the newline convention in the
86: \fIoptions\fP arguments for the compiling and matching functions.
87: .P
88: It is also possible to specify a newline convention by starting a pattern
89: string with one of the following five sequences:
90: .sp
91: (*CR) carriage return
92: (*LF) linefeed
93: (*CRLF) carriage return, followed by linefeed
94: (*ANYCRLF) any of the three above
95: (*ANY) all Unicode newline sequences
96: .sp
97: These override the default and the options given to \fBpcre_compile()\fP or
98: \fBpcre_compile2()\fP. For example, on a Unix system where LF is the default
99: newline sequence, the pattern
100: .sp
101: (*CR)a.b
102: .sp
103: changes the convention to CR. That pattern matches "a\enb" because LF is no
104: longer a newline. Note that these special settings, which are not
105: Perl-compatible, are recognized only at the very start of a pattern, and that
106: they must be in upper case. If more than one of them is present, the last one
107: is used.
108: .P
109: The newline convention affects the interpretation of the dot metacharacter when
110: PCRE_DOTALL is not set, and also the behaviour of \eN. However, it does not
111: affect what the \eR escape sequence matches. By default, this is any Unicode
112: newline sequence, for Perl compatibility. However, this can be changed; see the
113: description of \eR in the section entitled
114: .\" HTML <a href="#newlineseq">
115: .\" </a>
116: "Newline sequences"
117: .\"
118: below. A change of \eR setting can be combined with a change of newline
119: convention.
120: .
121: .
122: .SH "CHARACTERS AND METACHARACTERS"
123: .rs
124: .sp
125: A regular expression is a pattern that is matched against a subject string from
126: left to right. Most characters stand for themselves in a pattern, and match the
127: corresponding characters in the subject. As a trivial example, the pattern
128: .sp
129: The quick brown fox
130: .sp
131: matches a portion of a subject string that is identical to itself. When
132: caseless matching is specified (the PCRE_CASELESS option), letters are matched
133: independently of case. In UTF-8 mode, PCRE always understands the concept of
134: case for characters whose values are less than 128, so caseless matching is
135: always possible. For characters with higher values, the concept of case is
136: supported if PCRE is compiled with Unicode property support, but not otherwise.
137: If you want to use caseless matching for characters 128 and above, you must
138: ensure that PCRE is compiled with Unicode property support as well as with
139: UTF-8 support.
140: .P
141: The power of regular expressions comes from the ability to include alternatives
142: and repetitions in the pattern. These are encoded in the pattern by the use of
143: \fImetacharacters\fP, which do not stand for themselves but instead are
144: interpreted in some special way.
145: .P
146: There are two different sets of metacharacters: those that are recognized
147: anywhere in the pattern except within square brackets, and those that are
148: recognized within square brackets. Outside square brackets, the metacharacters
149: are as follows:
150: .sp
151: \e general escape character with several uses
152: ^ assert start of string (or line, in multiline mode)
153: $ assert end of string (or line, in multiline mode)
154: . match any character except newline (by default)
155: [ start character class definition
156: | start of alternative branch
157: ( start subpattern
158: ) end subpattern
159: ? extends the meaning of (
160: also 0 or 1 quantifier
161: also quantifier minimizer
162: * 0 or more quantifier
163: + 1 or more quantifier
164: also "possessive quantifier"
165: { start min/max quantifier
166: .sp
167: Part of a pattern that is in square brackets is called a "character class". In
168: a character class the only metacharacters are:
169: .sp
170: \e general escape character
171: ^ negate the class, but only if the first character
172: - indicates character range
173: .\" JOIN
174: [ POSIX character class (only if followed by POSIX
175: syntax)
176: ] terminates the character class
177: .sp
178: The following sections describe the use of each of the metacharacters.
179: .
180: .
181: .SH BACKSLASH
182: .rs
183: .sp
184: The backslash character has several uses. Firstly, if it is followed by a
185: character that is not a number or a letter, it takes away any special meaning
186: that character may have. This use of backslash as an escape character applies
187: both inside and outside character classes.
188: .P
189: For example, if you want to match a * character, you write \e* in the pattern.
190: This escaping action applies whether or not the following character would
191: otherwise be interpreted as a metacharacter, so it is always safe to precede a
192: non-alphanumeric with backslash to specify that it stands for itself. In
193: particular, if you want to match a backslash, you write \e\e.
194: .P
195: In UTF-8 mode, only ASCII numbers and letters have any special meaning after a
196: backslash. All other characters (in particular, those whose codepoints are
197: greater than 127) are treated as literals.
198: .P
199: If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
200: pattern (other than in a character class) and characters between a # outside
201: a character class and the next newline are ignored. An escaping backslash can
202: be used to include a whitespace or # character as part of the pattern.
203: .P
204: If you want to remove the special meaning from a sequence of characters, you
205: can do so by putting them between \eQ and \eE. This is different from Perl in
206: that $ and @ are handled as literals in \eQ...\eE sequences in PCRE, whereas in
207: Perl, $ and @ cause variable interpolation. Note the following examples:
208: .sp
209: Pattern PCRE matches Perl matches
210: .sp
211: .\" JOIN
212: \eQabc$xyz\eE abc$xyz abc followed by the
213: contents of $xyz
214: \eQabc\e$xyz\eE abc\e$xyz abc\e$xyz
215: \eQabc\eE\e$\eQxyz\eE abc$xyz abc$xyz
216: .sp
217: The \eQ...\eE sequence is recognized both inside and outside character classes.
218: An isolated \eE that is not preceded by \eQ is ignored. If \eQ is not followed
219: by \eE later in the pattern, the literal interpretation continues to the end of
220: the pattern (that is, \eE is assumed at the end). If the isolated \eQ is inside
221: a character class, this causes an error, because the character class is not
222: terminated.
223: .
224: .
225: .\" HTML <a name="digitsafterbackslash"></a>
226: .SS "Non-printing characters"
227: .rs
228: .sp
229: A second use of backslash provides a way of encoding non-printing characters
230: in patterns in a visible manner. There is no restriction on the appearance of
231: non-printing characters, apart from the binary zero that terminates a pattern,
232: but when a pattern is being prepared by text editing, it is often easier to use
233: one of the following escape sequences than the binary character it represents:
234: .sp
235: \ea alarm, that is, the BEL character (hex 07)
236: \ecx "control-x", where x is any ASCII character
237: \ee escape (hex 1B)
238: \ef formfeed (hex 0C)
239: \en linefeed (hex 0A)
240: \er carriage return (hex 0D)
241: \et tab (hex 09)
242: \eddd character with octal code ddd, or back reference
243: \exhh character with hex code hh
244: \ex{hhh..} character with hex code hhh.. (non-JavaScript mode)
245: \euhhhh character with hex code hhhh (JavaScript mode only)
246: .sp
247: The precise effect of \ecx is as follows: if x is a lower case letter, it
248: is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
249: Thus \ecz becomes hex 1A (z is 7A), but \ec{ becomes hex 3B ({ is 7B), while
250: \ec; becomes hex 7B (; is 3B). If the byte following \ec has a value greater
251: than 127, a compile-time error occurs. This locks out non-ASCII characters in
252: both byte mode and UTF-8 mode. (When PCRE is compiled in EBCDIC mode, all byte
253: values are valid. A lower case letter is converted to upper case, and then the
254: 0xc0 bits are flipped.)
255: .P
256: By default, after \ex, from zero to two hexadecimal digits are read (letters
257: can be in upper or lower case). Any number of hexadecimal digits may appear
258: between \ex{ and }, but the value of the character code must be less than 256
259: in non-UTF-8 mode, and less than 2**31 in UTF-8 mode. That is, the maximum
260: value in hexadecimal is 7FFFFFFF. Note that this is bigger than the largest
261: Unicode code point, which is 10FFFF.
262: .P
263: If characters other than hexadecimal digits appear between \ex{ and }, or if
264: there is no terminating }, this form of escape is not recognized. Instead, the
265: initial \ex will be interpreted as a basic hexadecimal escape, with no
266: following digits, giving a character whose value is zero.
267: .P
268: If the PCRE_JAVASCRIPT_COMPAT option is set, the interpretation of \ex is
269: as just described only when it is followed by two hexadecimal digits.
270: Otherwise, it matches a literal "x" character. In JavaScript mode, support for
271: code points greater than 256 is provided by \eu, which must be followed by
272: four hexadecimal digits; otherwise it matches a literal "u" character.
273: .P
274: Characters whose value is less than 256 can be defined by either of the two
275: syntaxes for \ex (or by \eu in JavaScript mode). There is no difference in the
276: way they are handled. For example, \exdc is exactly the same as \ex{dc} (or
277: \eu00dc in JavaScript mode).
278: .P
279: After \e0 up to two further octal digits are read. If there are fewer than two
280: digits, just those that are present are used. Thus the sequence \e0\ex\e07
281: specifies two binary zeros followed by a BEL character (code value 7). Make
282: sure you supply two digits after the initial zero if the pattern character that
283: follows is itself an octal digit.
284: .P
285: The handling of a backslash followed by a digit other than 0 is complicated.
286: Outside a character class, PCRE reads it and any following digits as a decimal
287: number. If the number is less than 10, or if there have been at least that many
288: previous capturing left parentheses in the expression, the entire sequence is
289: taken as a \fIback reference\fP. A description of how this works is given
290: .\" HTML <a href="#backreferences">
291: .\" </a>
292: later,
293: .\"
294: following the discussion of
295: .\" HTML <a href="#subpattern">
296: .\" </a>
297: parenthesized subpatterns.
298: .\"
299: .P
300: Inside a character class, or if the decimal number is greater than 9 and there
301: have not been that many capturing subpatterns, PCRE re-reads up to three octal
302: digits following the backslash, and uses them to generate a data character. Any
303: subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
304: character specified in octal must be less than \e400. In UTF-8 mode, values up
305: to \e777 are permitted. For example:
306: .sp
307: \e040 is another way of writing a space
308: .\" JOIN
309: \e40 is the same, provided there are fewer than 40
310: previous capturing subpatterns
311: \e7 is always a back reference
312: .\" JOIN
313: \e11 might be a back reference, or another way of
314: writing a tab
315: \e011 is always a tab
316: \e0113 is a tab followed by the character "3"
317: .\" JOIN
318: \e113 might be a back reference, otherwise the
319: character with octal code 113
320: .\" JOIN
321: \e377 might be a back reference, otherwise
322: the byte consisting entirely of 1 bits
323: .\" JOIN
324: \e81 is either a back reference, or a binary zero
325: followed by the two characters "8" and "1"
326: .sp
327: Note that octal values of 100 or greater must not be introduced by a leading
328: zero, because no more than three octal digits are ever read.
329: .P
330: All the sequences that define a single character value can be used both inside
331: and outside character classes. In addition, inside a character class, \eb is
332: interpreted as the backspace character (hex 08).
333: .P
334: \eN is not allowed in a character class. \eB, \eR, and \eX are not special
335: inside a character class. Like other unrecognized escape sequences, they are
336: treated as the literal characters "B", "R", and "X" by default, but cause an
337: error if the PCRE_EXTRA option is set. Outside a character class, these
338: sequences have different meanings.
339: .
340: .
341: .SS "Unsupported escape sequences"
342: .rs
343: .sp
344: In Perl, the sequences \el, \eL, \eu, and \eU are recognized by its string
345: handler and used to modify the case of following characters. By default, PCRE
346: does not support these escape sequences. However, if the PCRE_JAVASCRIPT_COMPAT
347: option is set, \eU matches a "U" character, and \eu can be used to define a
348: character by code point, as described in the previous section.
349: .
350: .
351: .SS "Absolute and relative back references"
352: .rs
353: .sp
354: The sequence \eg followed by an unsigned or a negative number, optionally
355: enclosed in braces, is an absolute or relative back reference. A named back
356: reference can be coded as \eg{name}. Back references are discussed
357: .\" HTML <a href="#backreferences">
358: .\" </a>
359: later,
360: .\"
361: following the discussion of
362: .\" HTML <a href="#subpattern">
363: .\" </a>
364: parenthesized subpatterns.
365: .\"
366: .
367: .
368: .SS "Absolute and relative subroutine calls"
369: .rs
370: .sp
371: For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
372: a number enclosed either in angle brackets or single quotes, is an alternative
373: syntax for referencing a subpattern as a "subroutine". Details are discussed
374: .\" HTML <a href="#onigurumasubroutines">
375: .\" </a>
376: later.
377: .\"
378: Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
379: synonymous. The former is a back reference; the latter is a
380: .\" HTML <a href="#subpatternsassubroutines">
381: .\" </a>
382: subroutine
383: .\"
384: call.
385: .
386: .
387: .\" HTML <a name="genericchartypes"></a>
388: .SS "Generic character types"
389: .rs
390: .sp
391: Another use of backslash is for specifying generic character types:
392: .sp
393: \ed any decimal digit
394: \eD any character that is not a decimal digit
395: \eh any horizontal whitespace character
396: \eH any character that is not a horizontal whitespace character
397: \es any whitespace character
398: \eS any character that is not a whitespace character
399: \ev any vertical whitespace character
400: \eV any character that is not a vertical whitespace character
401: \ew any "word" character
402: \eW any "non-word" character
403: .sp
404: There is also the single sequence \eN, which matches a non-newline character.
405: This is the same as
406: .\" HTML <a href="#fullstopdot">
407: .\" </a>
408: the "." metacharacter
409: .\"
410: when PCRE_DOTALL is not set. Perl also uses \eN to match characters by name;
411: PCRE does not support this.
412: .P
413: Each pair of lower and upper case escape sequences partitions the complete set
414: of characters into two disjoint sets. Any given character matches one, and only
415: one, of each pair. The sequences can appear both inside and outside character
416: classes. They each match one character of the appropriate type. If the current
417: matching point is at the end of the subject string, all of them fail, because
418: there is no character to match.
419: .P
420: For compatibility with Perl, \es does not match the VT character (code 11).
421: This makes it different from the the POSIX "space" class. The \es characters
422: are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
423: included in a Perl script, \es may match the VT character. In PCRE, it never
424: does.
425: .P
426: A "word" character is an underscore or any character that is a letter or digit.
427: By default, the definition of letters and digits is controlled by PCRE's
428: low-valued character tables, and may vary if locale-specific matching is taking
429: place (see
430: .\" HTML <a href="pcreapi.html#localesupport">
431: .\" </a>
432: "Locale support"
433: .\"
434: in the
435: .\" HREF
436: \fBpcreapi\fP
437: .\"
438: page). For example, in a French locale such as "fr_FR" in Unix-like systems,
439: or "french" in Windows, some character codes greater than 128 are used for
440: accented letters, and these are then matched by \ew. The use of locales with
441: Unicode is discouraged.
442: .P
443: By default, in UTF-8 mode, characters with values greater than 128 never match
444: \ed, \es, or \ew, and always match \eD, \eS, and \eW. These sequences retain
445: their original meanings from before UTF-8 support was available, mainly for
446: efficiency reasons. However, if PCRE is compiled with Unicode property support,
447: and the PCRE_UCP option is set, the behaviour is changed so that Unicode
448: properties are used to determine character types, as follows:
449: .sp
450: \ed any character that \ep{Nd} matches (decimal digit)
451: \es any character that \ep{Z} matches, plus HT, LF, FF, CR
452: \ew any character that \ep{L} or \ep{N} matches, plus underscore
453: .sp
454: The upper case escapes match the inverse sets of characters. Note that \ed
455: matches only decimal digits, whereas \ew matches any Unicode digit, as well as
456: any Unicode letter, and underscore. Note also that PCRE_UCP affects \eb, and
457: \eB because they are defined in terms of \ew and \eW. Matching these sequences
458: is noticeably slower when PCRE_UCP is set.
459: .P
460: The sequences \eh, \eH, \ev, and \eV are features that were added to Perl at
461: release 5.10. In contrast to the other sequences, which match only ASCII
462: characters by default, these always match certain high-valued codepoints in
463: UTF-8 mode, whether or not PCRE_UCP is set. The horizontal space characters
464: are:
465: .sp
466: U+0009 Horizontal tab
467: U+0020 Space
468: U+00A0 Non-break space
469: U+1680 Ogham space mark
470: U+180E Mongolian vowel separator
471: U+2000 En quad
472: U+2001 Em quad
473: U+2002 En space
474: U+2003 Em space
475: U+2004 Three-per-em space
476: U+2005 Four-per-em space
477: U+2006 Six-per-em space
478: U+2007 Figure space
479: U+2008 Punctuation space
480: U+2009 Thin space
481: U+200A Hair space
482: U+202F Narrow no-break space
483: U+205F Medium mathematical space
484: U+3000 Ideographic space
485: .sp
486: The vertical space characters are:
487: .sp
488: U+000A Linefeed
489: U+000B Vertical tab
490: U+000C Formfeed
491: U+000D Carriage return
492: U+0085 Next line
493: U+2028 Line separator
494: U+2029 Paragraph separator
495: .
496: .
497: .\" HTML <a name="newlineseq"></a>
498: .SS "Newline sequences"
499: .rs
500: .sp
501: Outside a character class, by default, the escape sequence \eR matches any
502: Unicode newline sequence. In non-UTF-8 mode \eR is equivalent to the following:
503: .sp
504: (?>\er\en|\en|\ex0b|\ef|\er|\ex85)
505: .sp
506: This is an example of an "atomic group", details of which are given
507: .\" HTML <a href="#atomicgroup">
508: .\" </a>
509: below.
510: .\"
511: This particular group matches either the two-character sequence CR followed by
512: LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
513: U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
514: line, U+0085). The two-character sequence is treated as a single unit that
515: cannot be split.
516: .P
517: In UTF-8 mode, two additional characters whose codepoints are greater than 255
518: are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
519: Unicode character property support is not needed for these characters to be
520: recognized.
521: .P
522: It is possible to restrict \eR to match only CR, LF, or CRLF (instead of the
523: complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
524: either at compile time or when the pattern is matched. (BSR is an abbrevation
525: for "backslash R".) This can be made the default when PCRE is built; if this is
526: the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
527: It is also possible to specify these settings by starting a pattern string with
528: one of the following sequences:
529: .sp
530: (*BSR_ANYCRLF) CR, LF, or CRLF only
531: (*BSR_UNICODE) any Unicode newline sequence
532: .sp
533: These override the default and the options given to \fBpcre_compile()\fP or
534: \fBpcre_compile2()\fP, but they can be overridden by options given to
535: \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP. Note that these special settings,
536: which are not Perl-compatible, are recognized only at the very start of a
537: pattern, and that they must be in upper case. If more than one of them is
538: present, the last one is used. They can be combined with a change of newline
539: convention; for example, a pattern can start with:
540: .sp
541: (*ANY)(*BSR_ANYCRLF)
542: .sp
543: They can also be combined with the (*UTF8) or (*UCP) special sequences. Inside
544: a character class, \eR is treated as an unrecognized escape sequence, and so
545: matches the letter "R" by default, but causes an error if PCRE_EXTRA is set.
546: .
547: .
548: .\" HTML <a name="uniextseq"></a>
549: .SS Unicode character properties
550: .rs
551: .sp
552: When PCRE is built with Unicode character property support, three additional
553: escape sequences that match characters with specific properties are available.
554: When not in UTF-8 mode, these sequences are of course limited to testing
555: characters whose codepoints are less than 256, but they do work in this mode.
556: The extra escape sequences are:
557: .sp
558: \ep{\fIxx\fP} a character with the \fIxx\fP property
559: \eP{\fIxx\fP} a character without the \fIxx\fP property
560: \eX an extended Unicode sequence
561: .sp
562: The property names represented by \fIxx\fP above are limited to the Unicode
563: script names, the general category properties, "Any", which matches any
564: character (including newline), and some special PCRE properties (described
565: in the
566: .\" HTML <a href="#extraprops">
567: .\" </a>
568: next section).
569: .\"
570: Other Perl properties such as "InMusicalSymbols" are not currently supported by
571: PCRE. Note that \eP{Any} does not match any characters, so always causes a
572: match failure.
573: .P
574: Sets of Unicode characters are defined as belonging to certain scripts. A
575: character from one of these sets can be matched using a script name. For
576: example:
577: .sp
578: \ep{Greek}
579: \eP{Han}
580: .sp
581: Those that are not part of an identified script are lumped together as
582: "Common". The current list of scripts is:
583: .P
584: Arabic,
585: Armenian,
586: Avestan,
587: Balinese,
588: Bamum,
589: Bengali,
590: Bopomofo,
591: Braille,
592: Buginese,
593: Buhid,
594: Canadian_Aboriginal,
595: Carian,
596: Cham,
597: Cherokee,
598: Common,
599: Coptic,
600: Cuneiform,
601: Cypriot,
602: Cyrillic,
603: Deseret,
604: Devanagari,
605: Egyptian_Hieroglyphs,
606: Ethiopic,
607: Georgian,
608: Glagolitic,
609: Gothic,
610: Greek,
611: Gujarati,
612: Gurmukhi,
613: Han,
614: Hangul,
615: Hanunoo,
616: Hebrew,
617: Hiragana,
618: Imperial_Aramaic,
619: Inherited,
620: Inscriptional_Pahlavi,
621: Inscriptional_Parthian,
622: Javanese,
623: Kaithi,
624: Kannada,
625: Katakana,
626: Kayah_Li,
627: Kharoshthi,
628: Khmer,
629: Lao,
630: Latin,
631: Lepcha,
632: Limbu,
633: Linear_B,
634: Lisu,
635: Lycian,
636: Lydian,
637: Malayalam,
638: Meetei_Mayek,
639: Mongolian,
640: Myanmar,
641: New_Tai_Lue,
642: Nko,
643: Ogham,
644: Old_Italic,
645: Old_Persian,
646: Old_South_Arabian,
647: Old_Turkic,
648: Ol_Chiki,
649: Oriya,
650: Osmanya,
651: Phags_Pa,
652: Phoenician,
653: Rejang,
654: Runic,
655: Samaritan,
656: Saurashtra,
657: Shavian,
658: Sinhala,
659: Sundanese,
660: Syloti_Nagri,
661: Syriac,
662: Tagalog,
663: Tagbanwa,
664: Tai_Le,
665: Tai_Tham,
666: Tai_Viet,
667: Tamil,
668: Telugu,
669: Thaana,
670: Thai,
671: Tibetan,
672: Tifinagh,
673: Ugaritic,
674: Vai,
675: Yi.
676: .P
677: Each character has exactly one Unicode general category property, specified by
678: a two-letter abbreviation. For compatibility with Perl, negation can be
679: specified by including a circumflex between the opening brace and the property
680: name. For example, \ep{^Lu} is the same as \eP{Lu}.
681: .P
682: If only one letter is specified with \ep or \eP, it includes all the general
683: category properties that start with that letter. In this case, in the absence
684: of negation, the curly brackets in the escape sequence are optional; these two
685: examples have the same effect:
686: .sp
687: \ep{L}
688: \epL
689: .sp
690: The following general category property codes are supported:
691: .sp
692: C Other
693: Cc Control
694: Cf Format
695: Cn Unassigned
696: Co Private use
697: Cs Surrogate
698: .sp
699: L Letter
700: Ll Lower case letter
701: Lm Modifier letter
702: Lo Other letter
703: Lt Title case letter
704: Lu Upper case letter
705: .sp
706: M Mark
707: Mc Spacing mark
708: Me Enclosing mark
709: Mn Non-spacing mark
710: .sp
711: N Number
712: Nd Decimal number
713: Nl Letter number
714: No Other number
715: .sp
716: P Punctuation
717: Pc Connector punctuation
718: Pd Dash punctuation
719: Pe Close punctuation
720: Pf Final punctuation
721: Pi Initial punctuation
722: Po Other punctuation
723: Ps Open punctuation
724: .sp
725: S Symbol
726: Sc Currency symbol
727: Sk Modifier symbol
728: Sm Mathematical symbol
729: So Other symbol
730: .sp
731: Z Separator
732: Zl Line separator
733: Zp Paragraph separator
734: Zs Space separator
735: .sp
736: The special property L& is also supported: it matches a character that has
737: the Lu, Ll, or Lt property, in other words, a letter that is not classified as
738: a modifier or "other".
739: .P
740: The Cs (Surrogate) property applies only to characters in the range U+D800 to
741: U+DFFF. Such characters are not valid in UTF-8 strings (see RFC 3629) and so
742: cannot be tested by PCRE, unless UTF-8 validity checking has been turned off
743: (see the discussion of PCRE_NO_UTF8_CHECK in the
744: .\" HREF
745: \fBpcreapi\fP
746: .\"
747: page). Perl does not support the Cs property.
748: .P
749: The long synonyms for property names that Perl supports (such as \ep{Letter})
750: are not supported by PCRE, nor is it permitted to prefix any of these
751: properties with "Is".
752: .P
753: No character that is in the Unicode table has the Cn (unassigned) property.
754: Instead, this property is assumed for any code point that is not in the
755: Unicode table.
756: .P
757: Specifying caseless matching does not affect these escape sequences. For
758: example, \ep{Lu} always matches only upper case letters.
759: .P
760: The \eX escape matches any number of Unicode characters that form an extended
761: Unicode sequence. \eX is equivalent to
762: .sp
763: (?>\ePM\epM*)
764: .sp
765: That is, it matches a character without the "mark" property, followed by zero
766: or more characters with the "mark" property, and treats the sequence as an
767: atomic group
768: .\" HTML <a href="#atomicgroup">
769: .\" </a>
770: (see below).
771: .\"
772: Characters with the "mark" property are typically accents that affect the
773: preceding character. None of them have codepoints less than 256, so in
774: non-UTF-8 mode \eX matches any one character.
775: .P
776: Note that recent versions of Perl have changed \eX to match what Unicode calls
777: an "extended grapheme cluster", which has a more complicated definition.
778: .P
779: Matching characters by Unicode property is not fast, because PCRE has to search
780: a structure that contains data for over fifteen thousand characters. That is
781: why the traditional escape sequences such as \ed and \ew do not use Unicode
782: properties in PCRE by default, though you can make them do so by setting the
783: PCRE_UCP option for \fBpcre_compile()\fP or by starting the pattern with
784: (*UCP).
785: .
786: .
787: .\" HTML <a name="extraprops"></a>
788: .SS PCRE's additional properties
789: .rs
790: .sp
791: As well as the standard Unicode properties described in the previous
792: section, PCRE supports four more that make it possible to convert traditional
793: escape sequences such as \ew and \es and POSIX character classes to use Unicode
794: properties. PCRE uses these non-standard, non-Perl properties internally when
795: PCRE_UCP is set. They are:
796: .sp
797: Xan Any alphanumeric character
798: Xps Any POSIX space character
799: Xsp Any Perl space character
800: Xwd Any Perl "word" character
801: .sp
802: Xan matches characters that have either the L (letter) or the N (number)
803: property. Xps matches the characters tab, linefeed, vertical tab, formfeed, or
804: carriage return, and any other character that has the Z (separator) property.
805: Xsp is the same as Xps, except that vertical tab is excluded. Xwd matches the
806: same characters as Xan, plus underscore.
807: .
808: .
809: .\" HTML <a name="resetmatchstart"></a>
810: .SS "Resetting the match start"
811: .rs
812: .sp
813: The escape sequence \eK causes any previously matched characters not to be
814: included in the final matched sequence. For example, the pattern:
815: .sp
816: foo\eKbar
817: .sp
818: matches "foobar", but reports that it has matched "bar". This feature is
819: similar to a lookbehind assertion
820: .\" HTML <a href="#lookbehind">
821: .\" </a>
822: (described below).
823: .\"
824: However, in this case, the part of the subject before the real match does not
825: have to be of fixed length, as lookbehind assertions do. The use of \eK does
826: not interfere with the setting of
827: .\" HTML <a href="#subpattern">
828: .\" </a>
829: captured substrings.
830: .\"
831: For example, when the pattern
832: .sp
833: (foo)\eKbar
834: .sp
835: matches "foobar", the first substring is still set to "foo".
836: .P
837: Perl documents that the use of \eK within assertions is "not well defined". In
838: PCRE, \eK is acted upon when it occurs inside positive assertions, but is
839: ignored in negative assertions.
840: .
841: .
842: .\" HTML <a name="smallassertions"></a>
843: .SS "Simple assertions"
844: .rs
845: .sp
846: The final use of backslash is for certain simple assertions. An assertion
847: specifies a condition that has to be met at a particular point in a match,
848: without consuming any characters from the subject string. The use of
849: subpatterns for more complicated assertions is described
850: .\" HTML <a href="#bigassertions">
851: .\" </a>
852: below.
853: .\"
854: The backslashed assertions are:
855: .sp
856: \eb matches at a word boundary
857: \eB matches when not at a word boundary
858: \eA matches at the start of the subject
859: \eZ matches at the end of the subject
860: also matches before a newline at the end of the subject
861: \ez matches only at the end of the subject
862: \eG matches at the first matching position in the subject
863: .sp
864: Inside a character class, \eb has a different meaning; it matches the backspace
865: character. If any other of these assertions appears in a character class, by
866: default it matches the corresponding literal character (for example, \eB
867: matches the letter B). However, if the PCRE_EXTRA option is set, an "invalid
868: escape sequence" error is generated instead.
869: .P
870: A word boundary is a position in the subject string where the current character
871: and the previous character do not both match \ew or \eW (i.e. one matches
872: \ew and the other matches \eW), or the start or end of the string if the
873: first or last character matches \ew, respectively. In UTF-8 mode, the meanings
874: of \ew and \eW can be changed by setting the PCRE_UCP option. When this is
875: done, it also affects \eb and \eB. Neither PCRE nor Perl has a separate "start
876: of word" or "end of word" metasequence. However, whatever follows \eb normally
877: determines which it is. For example, the fragment \eba matches "a" at the start
878: of a word.
879: .P
880: The \eA, \eZ, and \ez assertions differ from the traditional circumflex and
881: dollar (described in the next section) in that they only ever match at the very
882: start and end of the subject string, whatever options are set. Thus, they are
883: independent of multiline mode. These three assertions are not affected by the
884: PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
885: circumflex and dollar metacharacters. However, if the \fIstartoffset\fP
886: argument of \fBpcre_exec()\fP is non-zero, indicating that matching is to start
887: at a point other than the beginning of the subject, \eA can never match. The
888: difference between \eZ and \ez is that \eZ matches before a newline at the end
889: of the string as well as at the very end, whereas \ez matches only at the end.
890: .P
891: The \eG assertion is true only when the current matching position is at the
892: start point of the match, as specified by the \fIstartoffset\fP argument of
893: \fBpcre_exec()\fP. It differs from \eA when the value of \fIstartoffset\fP is
894: non-zero. By calling \fBpcre_exec()\fP multiple times with appropriate
895: arguments, you can mimic Perl's /g option, and it is in this kind of
896: implementation where \eG can be useful.
897: .P
898: Note, however, that PCRE's interpretation of \eG, as the start of the current
899: match, is subtly different from Perl's, which defines it as the end of the
900: previous match. In Perl, these can be different when the previously matched
901: string was empty. Because PCRE does just one match at a time, it cannot
902: reproduce this behaviour.
903: .P
904: If all the alternatives of a pattern begin with \eG, the expression is anchored
905: to the starting match position, and the "anchored" flag is set in the compiled
906: regular expression.
907: .
908: .
909: .SH "CIRCUMFLEX AND DOLLAR"
910: .rs
911: .sp
912: Outside a character class, in the default matching mode, the circumflex
913: character is an assertion that is true only if the current matching point is
914: at the start of the subject string. If the \fIstartoffset\fP argument of
915: \fBpcre_exec()\fP is non-zero, circumflex can never match if the PCRE_MULTILINE
916: option is unset. Inside a character class, circumflex has an entirely different
917: meaning
918: .\" HTML <a href="#characterclass">
919: .\" </a>
920: (see below).
921: .\"
922: .P
923: Circumflex need not be the first character of the pattern if a number of
924: alternatives are involved, but it should be the first thing in each alternative
925: in which it appears if the pattern is ever to match that branch. If all
926: possible alternatives start with a circumflex, that is, if the pattern is
927: constrained to match only at the start of the subject, it is said to be an
928: "anchored" pattern. (There are also other constructs that can cause a pattern
929: to be anchored.)
930: .P
931: A dollar character is an assertion that is true only if the current matching
932: point is at the end of the subject string, or immediately before a newline
933: at the end of the string (by default). Dollar need not be the last character of
934: the pattern if a number of alternatives are involved, but it should be the last
935: item in any branch in which it appears. Dollar has no special meaning in a
936: character class.
937: .P
938: The meaning of dollar can be changed so that it matches only at the very end of
939: the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
940: does not affect the \eZ assertion.
941: .P
942: The meanings of the circumflex and dollar characters are changed if the
943: PCRE_MULTILINE option is set. When this is the case, a circumflex matches
944: immediately after internal newlines as well as at the start of the subject
945: string. It does not match after a newline that ends the string. A dollar
946: matches before any newlines in the string, as well as at the very end, when
947: PCRE_MULTILINE is set. When newline is specified as the two-character
948: sequence CRLF, isolated CR and LF characters do not indicate newlines.
949: .P
950: For example, the pattern /^abc$/ matches the subject string "def\enabc" (where
951: \en represents a newline) in multiline mode, but not otherwise. Consequently,
952: patterns that are anchored in single line mode because all branches start with
953: ^ are not anchored in multiline mode, and a match for circumflex is possible
954: when the \fIstartoffset\fP argument of \fBpcre_exec()\fP is non-zero. The
955: PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
956: .P
957: Note that the sequences \eA, \eZ, and \ez can be used to match the start and
958: end of the subject in both modes, and if all branches of a pattern start with
959: \eA it is always anchored, whether or not PCRE_MULTILINE is set.
960: .
961: .
962: .\" HTML <a name="fullstopdot"></a>
963: .SH "FULL STOP (PERIOD, DOT) AND \eN"
964: .rs
965: .sp
966: Outside a character class, a dot in the pattern matches any one character in
967: the subject string except (by default) a character that signifies the end of a
968: line. In UTF-8 mode, the matched character may be more than one byte long.
969: .P
970: When a line ending is defined as a single character, dot never matches that
971: character; when the two-character sequence CRLF is used, dot does not match CR
972: if it is immediately followed by LF, but otherwise it matches all characters
973: (including isolated CRs and LFs). When any Unicode line endings are being
974: recognized, dot does not match CR or LF or any of the other line ending
975: characters.
976: .P
977: The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
978: option is set, a dot matches any one character, without exception. If the
979: two-character sequence CRLF is present in the subject string, it takes two dots
980: to match it.
981: .P
982: The handling of dot is entirely independent of the handling of circumflex and
983: dollar, the only relationship being that they both involve newlines. Dot has no
984: special meaning in a character class.
985: .P
986: The escape sequence \eN behaves like a dot, except that it is not affected by
987: the PCRE_DOTALL option. In other words, it matches any character except one
988: that signifies the end of a line. Perl also uses \eN to match characters by
989: name; PCRE does not support this.
990: .
991: .
992: .SH "MATCHING A SINGLE BYTE"
993: .rs
994: .sp
995: Outside a character class, the escape sequence \eC matches any one byte, both
996: in and out of UTF-8 mode. Unlike a dot, it always matches line-ending
997: characters. The feature is provided in Perl in order to match individual bytes
998: in UTF-8 mode, but it is unclear how it can usefully be used. Because \eC
999: breaks up characters into individual bytes, matching one byte with \eC in UTF-8
1000: mode means that the rest of the string may start with a malformed UTF-8
1001: character. This has undefined results, because PCRE assumes that it is dealing
1002: with valid UTF-8 strings (and by default it checks this at the start of
1003: processing unless the PCRE_NO_UTF8_CHECK option is used).
1004: .P
1005: PCRE does not allow \eC to appear in lookbehind assertions
1006: .\" HTML <a href="#lookbehind">
1007: .\" </a>
1008: (described below)
1009: .\"
1010: in UTF-8 mode, because this would make it impossible to calculate the length of
1011: the lookbehind.
1012: .P
1013: In general, the \eC escape sequence is best avoided in UTF-8 mode. However, one
1014: way of using it that avoids the problem of malformed UTF-8 characters is to
1015: use a lookahead to check the length of the next character, as in this pattern
1016: (ignore white space and line breaks):
1017: .sp
1018: (?| (?=[\ex00-\ex7f])(\eC) |
1019: (?=[\ex80-\ex{7ff}])(\eC)(\eC) |
1020: (?=[\ex{800}-\ex{ffff}])(\eC)(\eC)(\eC) |
1021: (?=[\ex{10000}-\ex{1fffff}])(\eC)(\eC)(\eC)(\eC))
1022: .sp
1023: A group that starts with (?| resets the capturing parentheses numbers in each
1024: alternative (see
1025: .\" HTML <a href="#dupsubpatternnumber">
1026: .\" </a>
1027: "Duplicate Subpattern Numbers"
1028: .\"
1029: below). The assertions at the start of each branch check the next UTF-8
1030: character for values whose encoding uses 1, 2, 3, or 4 bytes, respectively. The
1031: character's individual bytes are then captured by the appropriate number of
1032: groups.
1033: .
1034: .
1035: .\" HTML <a name="characterclass"></a>
1036: .SH "SQUARE BRACKETS AND CHARACTER CLASSES"
1037: .rs
1038: .sp
1039: An opening square bracket introduces a character class, terminated by a closing
1040: square bracket. A closing square bracket on its own is not special by default.
1041: However, if the PCRE_JAVASCRIPT_COMPAT option is set, a lone closing square
1042: bracket causes a compile-time error. If a closing square bracket is required as
1043: a member of the class, it should be the first data character in the class
1044: (after an initial circumflex, if present) or escaped with a backslash.
1045: .P
1046: A character class matches a single character in the subject. In UTF-8 mode, the
1047: character may be more than one byte long. A matched character must be in the
1048: set of characters defined by the class, unless the first character in the class
1049: definition is a circumflex, in which case the subject character must not be in
1050: the set defined by the class. If a circumflex is actually required as a member
1051: of the class, ensure it is not the first character, or escape it with a
1052: backslash.
1053: .P
1054: For example, the character class [aeiou] matches any lower case vowel, while
1055: [^aeiou] matches any character that is not a lower case vowel. Note that a
1056: circumflex is just a convenient notation for specifying the characters that
1057: are in the class by enumerating those that are not. A class that starts with a
1058: circumflex is not an assertion; it still consumes a character from the subject
1059: string, and therefore it fails if the current pointer is at the end of the
1060: string.
1061: .P
1062: In UTF-8 mode, characters with values greater than 255 can be included in a
1063: class as a literal string of bytes, or by using the \ex{ escaping mechanism.
1064: .P
1065: When caseless matching is set, any letters in a class represent both their
1066: upper case and lower case versions, so for example, a caseless [aeiou] matches
1067: "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
1068: caseful version would. In UTF-8 mode, PCRE always understands the concept of
1069: case for characters whose values are less than 128, so caseless matching is
1070: always possible. For characters with higher values, the concept of case is
1071: supported if PCRE is compiled with Unicode property support, but not otherwise.
1072: If you want to use caseless matching in UTF8-mode for characters 128 and above,
1073: you must ensure that PCRE is compiled with Unicode property support as well as
1074: with UTF-8 support.
1075: .P
1076: Characters that might indicate line breaks are never treated in any special way
1077: when matching character classes, whatever line-ending sequence is in use, and
1078: whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
1079: such as [^a] always matches one of these characters.
1080: .P
1081: The minus (hyphen) character can be used to specify a range of characters in a
1082: character class. For example, [d-m] matches any letter between d and m,
1083: inclusive. If a minus character is required in a class, it must be escaped with
1084: a backslash or appear in a position where it cannot be interpreted as
1085: indicating a range, typically as the first or last character in the class.
1086: .P
1087: It is not possible to have the literal character "]" as the end character of a
1088: range. A pattern such as [W-]46] is interpreted as a class of two characters
1089: ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
1090: "-46]". However, if the "]" is escaped with a backslash it is interpreted as
1091: the end of range, so [W-\e]46] is interpreted as a class containing a range
1092: followed by two other characters. The octal or hexadecimal representation of
1093: "]" can also be used to end a range.
1094: .P
1095: Ranges operate in the collating sequence of character values. They can also be
1096: used for characters specified numerically, for example [\e000-\e037]. In UTF-8
1097: mode, ranges can include characters whose values are greater than 255, for
1098: example [\ex{100}-\ex{2ff}].
1099: .P
1100: If a range that includes letters is used when caseless matching is set, it
1101: matches the letters in either case. For example, [W-c] is equivalent to
1102: [][\e\e^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
1103: tables for a French locale are in use, [\exc8-\excb] matches accented E
1104: characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
1105: characters with values greater than 128 only when it is compiled with Unicode
1106: property support.
1107: .P
1108: The character escape sequences \ed, \eD, \eh, \eH, \ep, \eP, \es, \eS, \ev,
1109: \eV, \ew, and \eW may appear in a character class, and add the characters that
1110: they match to the class. For example, [\edABCDEF] matches any hexadecimal
1111: digit. In UTF-8 mode, the PCRE_UCP option affects the meanings of \ed, \es, \ew
1112: and their upper case partners, just as it does when they appear outside a
1113: character class, as described in the section entitled
1114: .\" HTML <a href="#genericchartypes">
1115: .\" </a>
1116: "Generic character types"
1117: .\"
1118: above. The escape sequence \eb has a different meaning inside a character
1119: class; it matches the backspace character. The sequences \eB, \eN, \eR, and \eX
1120: are not special inside a character class. Like any other unrecognized escape
1121: sequences, they are treated as the literal characters "B", "N", "R", and "X" by
1122: default, but cause an error if the PCRE_EXTRA option is set.
1123: .P
1124: A circumflex can conveniently be used with the upper case character types to
1125: specify a more restricted set of characters than the matching lower case type.
1126: For example, the class [^\eW_] matches any letter or digit, but not underscore,
1127: whereas [\ew] includes underscore. A positive character class should be read as
1128: "something OR something OR ..." and a negative class as "NOT something AND NOT
1129: something AND NOT ...".
1130: .P
1131: The only metacharacters that are recognized in character classes are backslash,
1132: hyphen (only where it can be interpreted as specifying a range), circumflex
1133: (only at the start), opening square bracket (only when it can be interpreted as
1134: introducing a POSIX class name - see the next section), and the terminating
1135: closing square bracket. However, escaping other non-alphanumeric characters
1136: does no harm.
1137: .
1138: .
1139: .SH "POSIX CHARACTER CLASSES"
1140: .rs
1141: .sp
1142: Perl supports the POSIX notation for character classes. This uses names
1143: enclosed by [: and :] within the enclosing square brackets. PCRE also supports
1144: this notation. For example,
1145: .sp
1146: [01[:alpha:]%]
1147: .sp
1148: matches "0", "1", any alphabetic character, or "%". The supported class names
1149: are:
1150: .sp
1151: alnum letters and digits
1152: alpha letters
1153: ascii character codes 0 - 127
1154: blank space or tab only
1155: cntrl control characters
1156: digit decimal digits (same as \ed)
1157: graph printing characters, excluding space
1158: lower lower case letters
1159: print printing characters, including space
1160: punct printing characters, excluding letters and digits and space
1161: space white space (not quite the same as \es)
1162: upper upper case letters
1163: word "word" characters (same as \ew)
1164: xdigit hexadecimal digits
1165: .sp
1166: The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
1167: space (32). Notice that this list includes the VT character (code 11). This
1168: makes "space" different to \es, which does not include VT (for Perl
1169: compatibility).
1170: .P
1171: The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
1172: 5.8. Another Perl extension is negation, which is indicated by a ^ character
1173: after the colon. For example,
1174: .sp
1175: [12[:^digit:]]
1176: .sp
1177: matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
1178: syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
1179: supported, and an error is given if they are encountered.
1180: .P
1181: By default, in UTF-8 mode, characters with values greater than 128 do not match
1182: any of the POSIX character classes. However, if the PCRE_UCP option is passed
1183: to \fBpcre_compile()\fP, some of the classes are changed so that Unicode
1184: character properties are used. This is achieved by replacing the POSIX classes
1185: by other sequences, as follows:
1186: .sp
1187: [:alnum:] becomes \ep{Xan}
1188: [:alpha:] becomes \ep{L}
1189: [:blank:] becomes \eh
1190: [:digit:] becomes \ep{Nd}
1191: [:lower:] becomes \ep{Ll}
1192: [:space:] becomes \ep{Xps}
1193: [:upper:] becomes \ep{Lu}
1194: [:word:] becomes \ep{Xwd}
1195: .sp
1196: Negated versions, such as [:^alpha:] use \eP instead of \ep. The other POSIX
1197: classes are unchanged, and match only characters with code points less than
1198: 128.
1199: .
1200: .
1201: .SH "VERTICAL BAR"
1202: .rs
1203: .sp
1204: Vertical bar characters are used to separate alternative patterns. For example,
1205: the pattern
1206: .sp
1207: gilbert|sullivan
1208: .sp
1209: matches either "gilbert" or "sullivan". Any number of alternatives may appear,
1210: and an empty alternative is permitted (matching the empty string). The matching
1211: process tries each alternative in turn, from left to right, and the first one
1212: that succeeds is used. If the alternatives are within a subpattern
1213: .\" HTML <a href="#subpattern">
1214: .\" </a>
1215: (defined below),
1216: .\"
1217: "succeeds" means matching the rest of the main pattern as well as the
1218: alternative in the subpattern.
1219: .
1220: .
1221: .SH "INTERNAL OPTION SETTING"
1222: .rs
1223: .sp
1224: The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
1225: PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
1226: the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
1227: The option letters are
1228: .sp
1229: i for PCRE_CASELESS
1230: m for PCRE_MULTILINE
1231: s for PCRE_DOTALL
1232: x for PCRE_EXTENDED
1233: .sp
1234: For example, (?im) sets caseless, multiline matching. It is also possible to
1235: unset these options by preceding the letter with a hyphen, and a combined
1236: setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1237: PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1238: permitted. If a letter appears both before and after the hyphen, the option is
1239: unset.
1240: .P
1241: The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
1242: changed in the same way as the Perl-compatible options by using the characters
1243: J, U and X respectively.
1244: .P
1245: When one of these option changes occurs at top level (that is, not inside
1246: subpattern parentheses), the change applies to the remainder of the pattern
1247: that follows. If the change is placed right at the start of a pattern, PCRE
1248: extracts it into the global options (and it will therefore show up in data
1249: extracted by the \fBpcre_fullinfo()\fP function).
1250: .P
1251: An option change within a subpattern (see below for a description of
1252: subpatterns) affects only that part of the subpattern that follows it, so
1253: .sp
1254: (a(?i)b)c
1255: .sp
1256: matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1257: By this means, options can be made to have different settings in different
1258: parts of the pattern. Any changes made in one alternative do carry on
1259: into subsequent branches within the same subpattern. For example,
1260: .sp
1261: (a(?i)b|c)
1262: .sp
1263: matches "ab", "aB", "c", and "C", even though when matching "C" the first
1264: branch is abandoned before the option setting. This is because the effects of
1265: option settings happen at compile time. There would be some very weird
1266: behaviour otherwise.
1267: .P
1268: \fBNote:\fP There are other PCRE-specific options that can be set by the
1269: application when the compile or match functions are called. In some cases the
1270: pattern can contain special leading sequences such as (*CRLF) to override what
1271: the application has set or what has been defaulted. Details are given in the
1272: section entitled
1273: .\" HTML <a href="#newlineseq">
1274: .\" </a>
1275: "Newline sequences"
1276: .\"
1277: above. There are also the (*UTF8) and (*UCP) leading sequences that can be used
1278: to set UTF-8 and Unicode property modes; they are equivalent to setting the
1279: PCRE_UTF8 and the PCRE_UCP options, respectively.
1280: .
1281: .
1282: .\" HTML <a name="subpattern"></a>
1283: .SH SUBPATTERNS
1284: .rs
1285: .sp
1286: Subpatterns are delimited by parentheses (round brackets), which can be nested.
1287: Turning part of a pattern into a subpattern does two things:
1288: .sp
1289: 1. It localizes a set of alternatives. For example, the pattern
1290: .sp
1291: cat(aract|erpillar|)
1292: .sp
1293: matches "cataract", "caterpillar", or "cat". Without the parentheses, it would
1294: match "cataract", "erpillar" or an empty string.
1295: .sp
1296: 2. It sets up the subpattern as a capturing subpattern. This means that, when
1297: the whole pattern matches, that portion of the subject string that matched the
1298: subpattern is passed back to the caller via the \fIovector\fP argument of
1299: \fBpcre_exec()\fP. Opening parentheses are counted from left to right (starting
1300: from 1) to obtain numbers for the capturing subpatterns. For example, if the
1301: string "the red king" is matched against the pattern
1302: .sp
1303: the ((red|white) (king|queen))
1304: .sp
1305: the captured substrings are "red king", "red", and "king", and are numbered 1,
1306: 2, and 3, respectively.
1307: .P
1308: The fact that plain parentheses fulfil two functions is not always helpful.
1309: There are often times when a grouping subpattern is required without a
1310: capturing requirement. If an opening parenthesis is followed by a question mark
1311: and a colon, the subpattern does not do any capturing, and is not counted when
1312: computing the number of any subsequent capturing subpatterns. For example, if
1313: the string "the white queen" is matched against the pattern
1314: .sp
1315: the ((?:red|white) (king|queen))
1316: .sp
1317: the captured substrings are "white queen" and "queen", and are numbered 1 and
1318: 2. The maximum number of capturing subpatterns is 65535.
1319: .P
1320: As a convenient shorthand, if any option settings are required at the start of
1321: a non-capturing subpattern, the option letters may appear between the "?" and
1322: the ":". Thus the two patterns
1323: .sp
1324: (?i:saturday|sunday)
1325: (?:(?i)saturday|sunday)
1326: .sp
1327: match exactly the same set of strings. Because alternative branches are tried
1328: from left to right, and options are not reset until the end of the subpattern
1329: is reached, an option setting in one branch does affect subsequent branches, so
1330: the above patterns match "SUNDAY" as well as "Saturday".
1331: .
1332: .
1333: .\" HTML <a name="dupsubpatternnumber"></a>
1334: .SH "DUPLICATE SUBPATTERN NUMBERS"
1335: .rs
1336: .sp
1337: Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
1338: the same numbers for its capturing parentheses. Such a subpattern starts with
1339: (?| and is itself a non-capturing subpattern. For example, consider this
1340: pattern:
1341: .sp
1342: (?|(Sat)ur|(Sun))day
1343: .sp
1344: Because the two alternatives are inside a (?| group, both sets of capturing
1345: parentheses are numbered one. Thus, when the pattern matches, you can look
1346: at captured substring number one, whichever alternative matched. This construct
1347: is useful when you want to capture part, but not all, of one of a number of
1348: alternatives. Inside a (?| group, parentheses are numbered as usual, but the
1349: number is reset at the start of each branch. The numbers of any capturing
1350: parentheses that follow the subpattern start after the highest number used in
1351: any branch. The following example is taken from the Perl documentation. The
1352: numbers underneath show in which buffer the captured content will be stored.
1353: .sp
1354: # before ---------------branch-reset----------- after
1355: / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1356: # 1 2 2 3 2 3 4
1357: .sp
1358: A back reference to a numbered subpattern uses the most recent value that is
1359: set for that number by any subpattern. The following pattern matches "abcabc"
1360: or "defdef":
1361: .sp
1362: /(?|(abc)|(def))\e1/
1363: .sp
1364: In contrast, a subroutine call to a numbered subpattern always refers to the
1365: first one in the pattern with the given number. The following pattern matches
1366: "abcabc" or "defabc":
1367: .sp
1368: /(?|(abc)|(def))(?1)/
1369: .sp
1370: If a
1371: .\" HTML <a href="#conditions">
1372: .\" </a>
1373: condition test
1374: .\"
1375: for a subpattern's having matched refers to a non-unique number, the test is
1376: true if any of the subpatterns of that number have matched.
1377: .P
1378: An alternative approach to using this "branch reset" feature is to use
1379: duplicate named subpatterns, as described in the next section.
1380: .
1381: .
1382: .SH "NAMED SUBPATTERNS"
1383: .rs
1384: .sp
1385: Identifying capturing parentheses by number is simple, but it can be very hard
1386: to keep track of the numbers in complicated regular expressions. Furthermore,
1387: if an expression is modified, the numbers may change. To help with this
1388: difficulty, PCRE supports the naming of subpatterns. This feature was not
1389: added to Perl until release 5.10. Python had the feature earlier, and PCRE
1390: introduced it at release 4.0, using the Python syntax. PCRE now supports both
1391: the Perl and the Python syntax. Perl allows identically numbered subpatterns to
1392: have different names, but PCRE does not.
1393: .P
1394: In PCRE, a subpattern can be named in one of three ways: (?<name>...) or
1395: (?'name'...) as in Perl, or (?P<name>...) as in Python. References to capturing
1396: parentheses from other parts of the pattern, such as
1397: .\" HTML <a href="#backreferences">
1398: .\" </a>
1399: back references,
1400: .\"
1401: .\" HTML <a href="#recursion">
1402: .\" </a>
1403: recursion,
1404: .\"
1405: and
1406: .\" HTML <a href="#conditions">
1407: .\" </a>
1408: conditions,
1409: .\"
1410: can be made by name as well as by number.
1411: .P
1412: Names consist of up to 32 alphanumeric characters and underscores. Named
1413: capturing parentheses are still allocated numbers as well as names, exactly as
1414: if the names were not present. The PCRE API provides function calls for
1415: extracting the name-to-number translation table from a compiled pattern. There
1416: is also a convenience function for extracting a captured substring by name.
1417: .P
1418: By default, a name must be unique within a pattern, but it is possible to relax
1419: this constraint by setting the PCRE_DUPNAMES option at compile time. (Duplicate
1420: names are also always permitted for subpatterns with the same number, set up as
1421: described in the previous section.) Duplicate names can be useful for patterns
1422: where only one instance of the named parentheses can match. Suppose you want to
1423: match the name of a weekday, either as a 3-letter abbreviation or as the full
1424: name, and in both cases you want to extract the abbreviation. This pattern
1425: (ignoring the line breaks) does the job:
1426: .sp
1427: (?<DN>Mon|Fri|Sun)(?:day)?|
1428: (?<DN>Tue)(?:sday)?|
1429: (?<DN>Wed)(?:nesday)?|
1430: (?<DN>Thu)(?:rsday)?|
1431: (?<DN>Sat)(?:urday)?
1432: .sp
1433: There are five capturing substrings, but only one is ever set after a match.
1434: (An alternative way of solving this problem is to use a "branch reset"
1435: subpattern, as described in the previous section.)
1436: .P
1437: The convenience function for extracting the data by name returns the substring
1438: for the first (and in this example, the only) subpattern of that name that
1439: matched. This saves searching to find which numbered subpattern it was.
1440: .P
1441: If you make a back reference to a non-unique named subpattern from elsewhere in
1442: the pattern, the one that corresponds to the first occurrence of the name is
1443: used. In the absence of duplicate numbers (see the previous section) this is
1444: the one with the lowest number. If you use a named reference in a condition
1445: test (see the
1446: .\"
1447: .\" HTML <a href="#conditions">
1448: .\" </a>
1449: section about conditions
1450: .\"
1451: below), either to check whether a subpattern has matched, or to check for
1452: recursion, all subpatterns with the same name are tested. If the condition is
1453: true for any one of them, the overall condition is true. This is the same
1454: behaviour as testing by number. For further details of the interfaces for
1455: handling named subpatterns, see the
1456: .\" HREF
1457: \fBpcreapi\fP
1458: .\"
1459: documentation.
1460: .P
1461: \fBWarning:\fP You cannot use different names to distinguish between two
1462: subpatterns with the same number because PCRE uses only the numbers when
1463: matching. For this reason, an error is given at compile time if different names
1464: are given to subpatterns with the same number. However, you can give the same
1465: name to subpatterns with the same number, even when PCRE_DUPNAMES is not set.
1466: .
1467: .
1468: .SH REPETITION
1469: .rs
1470: .sp
1471: Repetition is specified by quantifiers, which can follow any of the following
1472: items:
1473: .sp
1474: a literal data character
1475: the dot metacharacter
1476: the \eC escape sequence
1477: the \eX escape sequence (in UTF-8 mode with Unicode properties)
1478: the \eR escape sequence
1479: an escape such as \ed or \epL that matches a single character
1480: a character class
1481: a back reference (see next section)
1482: a parenthesized subpattern (including assertions)
1483: a subroutine call to a subpattern (recursive or otherwise)
1484: .sp
1485: The general repetition quantifier specifies a minimum and maximum number of
1486: permitted matches, by giving the two numbers in curly brackets (braces),
1487: separated by a comma. The numbers must be less than 65536, and the first must
1488: be less than or equal to the second. For example:
1489: .sp
1490: z{2,4}
1491: .sp
1492: matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1493: character. If the second number is omitted, but the comma is present, there is
1494: no upper limit; if the second number and the comma are both omitted, the
1495: quantifier specifies an exact number of required matches. Thus
1496: .sp
1497: [aeiou]{3,}
1498: .sp
1499: matches at least 3 successive vowels, but may match many more, while
1500: .sp
1501: \ed{8}
1502: .sp
1503: matches exactly 8 digits. An opening curly bracket that appears in a position
1504: where a quantifier is not allowed, or one that does not match the syntax of a
1505: quantifier, is taken as a literal character. For example, {,6} is not a
1506: quantifier, but a literal string of four characters.
1507: .P
1508: In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
1509: bytes. Thus, for example, \ex{100}{2} matches two UTF-8 characters, each of
1510: which is represented by a two-byte sequence. Similarly, when Unicode property
1511: support is available, \eX{3} matches three Unicode extended sequences, each of
1512: which may be several bytes long (and they may be of different lengths).
1513: .P
1514: The quantifier {0} is permitted, causing the expression to behave as if the
1515: previous item and the quantifier were not present. This may be useful for
1516: subpatterns that are referenced as
1517: .\" HTML <a href="#subpatternsassubroutines">
1518: .\" </a>
1519: subroutines
1520: .\"
1521: from elsewhere in the pattern (but see also the section entitled
1522: .\" HTML <a href="#subdefine">
1523: .\" </a>
1524: "Defining subpatterns for use by reference only"
1525: .\"
1526: below). Items other than subpatterns that have a {0} quantifier are omitted
1527: from the compiled pattern.
1528: .P
1529: For convenience, the three most common quantifiers have single-character
1530: abbreviations:
1531: .sp
1532: * is equivalent to {0,}
1533: + is equivalent to {1,}
1534: ? is equivalent to {0,1}
1535: .sp
1536: It is possible to construct infinite loops by following a subpattern that can
1537: match no characters with a quantifier that has no upper limit, for example:
1538: .sp
1539: (a?)*
1540: .sp
1541: Earlier versions of Perl and PCRE used to give an error at compile time for
1542: such patterns. However, because there are cases where this can be useful, such
1543: patterns are now accepted, but if any repetition of the subpattern does in fact
1544: match no characters, the loop is forcibly broken.
1545: .P
1546: By default, the quantifiers are "greedy", that is, they match as much as
1547: possible (up to the maximum number of permitted times), without causing the
1548: rest of the pattern to fail. The classic example of where this gives problems
1549: is in trying to match comments in C programs. These appear between /* and */
1550: and within the comment, individual * and / characters may appear. An attempt to
1551: match C comments by applying the pattern
1552: .sp
1553: /\e*.*\e*/
1554: .sp
1555: to the string
1556: .sp
1557: /* first comment */ not comment /* second comment */
1558: .sp
1559: fails, because it matches the entire string owing to the greediness of the .*
1560: item.
1561: .P
1562: However, if a quantifier is followed by a question mark, it ceases to be
1563: greedy, and instead matches the minimum number of times possible, so the
1564: pattern
1565: .sp
1566: /\e*.*?\e*/
1567: .sp
1568: does the right thing with the C comments. The meaning of the various
1569: quantifiers is not otherwise changed, just the preferred number of matches.
1570: Do not confuse this use of question mark with its use as a quantifier in its
1571: own right. Because it has two uses, it can sometimes appear doubled, as in
1572: .sp
1573: \ed??\ed
1574: .sp
1575: which matches one digit by preference, but can match two if that is the only
1576: way the rest of the pattern matches.
1577: .P
1578: If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
1579: the quantifiers are not greedy by default, but individual ones can be made
1580: greedy by following them with a question mark. In other words, it inverts the
1581: default behaviour.
1582: .P
1583: When a parenthesized subpattern is quantified with a minimum repeat count that
1584: is greater than 1 or with a limited maximum, more memory is required for the
1585: compiled pattern, in proportion to the size of the minimum or maximum.
1586: .P
1587: If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1588: to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
1589: implicitly anchored, because whatever follows will be tried against every
1590: character position in the subject string, so there is no point in retrying the
1591: overall match at any position after the first. PCRE normally treats such a
1592: pattern as though it were preceded by \eA.
1593: .P
1594: In cases where it is known that the subject string contains no newlines, it is
1595: worth setting PCRE_DOTALL in order to obtain this optimization, or
1596: alternatively using ^ to indicate anchoring explicitly.
1597: .P
1598: However, there is one situation where the optimization cannot be used. When .*
1599: is inside capturing parentheses that are the subject of a back reference
1600: elsewhere in the pattern, a match at the start may fail where a later one
1601: succeeds. Consider, for example:
1602: .sp
1603: (.*)abc\e1
1604: .sp
1605: If the subject is "xyz123abc123" the match point is the fourth character. For
1606: this reason, such a pattern is not implicitly anchored.
1607: .P
1608: When a capturing subpattern is repeated, the value captured is the substring
1609: that matched the final iteration. For example, after
1610: .sp
1611: (tweedle[dume]{3}\es*)+
1612: .sp
1613: has matched "tweedledum tweedledee" the value of the captured substring is
1614: "tweedledee". However, if there are nested capturing subpatterns, the
1615: corresponding captured values may have been set in previous iterations. For
1616: example, after
1617: .sp
1618: /(a|(b))+/
1619: .sp
1620: matches "aba" the value of the second captured substring is "b".
1621: .
1622: .
1623: .\" HTML <a name="atomicgroup"></a>
1624: .SH "ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS"
1625: .rs
1626: .sp
1627: With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
1628: repetition, failure of what follows normally causes the repeated item to be
1629: re-evaluated to see if a different number of repeats allows the rest of the
1630: pattern to match. Sometimes it is useful to prevent this, either to change the
1631: nature of the match, or to cause it fail earlier than it otherwise might, when
1632: the author of the pattern knows there is no point in carrying on.
1633: .P
1634: Consider, for example, the pattern \ed+foo when applied to the subject line
1635: .sp
1636: 123456bar
1637: .sp
1638: After matching all 6 digits and then failing to match "foo", the normal
1639: action of the matcher is to try again with only 5 digits matching the \ed+
1640: item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
1641: (a term taken from Jeffrey Friedl's book) provides the means for specifying
1642: that once a subpattern has matched, it is not to be re-evaluated in this way.
1643: .P
1644: If we use atomic grouping for the previous example, the matcher gives up
1645: immediately on failing to match "foo" the first time. The notation is a kind of
1646: special parenthesis, starting with (?> as in this example:
1647: .sp
1648: (?>\ed+)foo
1649: .sp
1650: This kind of parenthesis "locks up" the part of the pattern it contains once
1651: it has matched, and a failure further into the pattern is prevented from
1652: backtracking into it. Backtracking past it to previous items, however, works as
1653: normal.
1654: .P
1655: An alternative description is that a subpattern of this type matches the string
1656: of characters that an identical standalone pattern would match, if anchored at
1657: the current point in the subject string.
1658: .P
1659: Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
1660: the above example can be thought of as a maximizing repeat that must swallow
1661: everything it can. So, while both \ed+ and \ed+? are prepared to adjust the
1662: number of digits they match in order to make the rest of the pattern match,
1663: (?>\ed+) can only match an entire sequence of digits.
1664: .P
1665: Atomic groups in general can of course contain arbitrarily complicated
1666: subpatterns, and can be nested. However, when the subpattern for an atomic
1667: group is just a single repeated item, as in the example above, a simpler
1668: notation, called a "possessive quantifier" can be used. This consists of an
1669: additional + character following a quantifier. Using this notation, the
1670: previous example can be rewritten as
1671: .sp
1672: \ed++foo
1673: .sp
1674: Note that a possessive quantifier can be used with an entire group, for
1675: example:
1676: .sp
1677: (abc|xyz){2,3}+
1678: .sp
1679: Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
1680: option is ignored. They are a convenient notation for the simpler forms of
1681: atomic group. However, there is no difference in the meaning of a possessive
1682: quantifier and the equivalent atomic group, though there may be a performance
1683: difference; possessive quantifiers should be slightly faster.
1684: .P
1685: The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
1686: Jeffrey Friedl originated the idea (and the name) in the first edition of his
1687: book. Mike McCloskey liked it, so implemented it when he built Sun's Java
1688: package, and PCRE copied it from there. It ultimately found its way into Perl
1689: at release 5.10.
1690: .P
1691: PCRE has an optimization that automatically "possessifies" certain simple
1692: pattern constructs. For example, the sequence A+B is treated as A++B because
1693: there is no point in backtracking into a sequence of A's when B must follow.
1694: .P
1695: When a pattern contains an unlimited repeat inside a subpattern that can itself
1696: be repeated an unlimited number of times, the use of an atomic group is the
1697: only way to avoid some failing matches taking a very long time indeed. The
1698: pattern
1699: .sp
1700: (\eD+|<\ed+>)*[!?]
1701: .sp
1702: matches an unlimited number of substrings that either consist of non-digits, or
1703: digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1704: quickly. However, if it is applied to
1705: .sp
1706: aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1707: .sp
1708: it takes a long time before reporting failure. This is because the string can
1709: be divided between the internal \eD+ repeat and the external * repeat in a
1710: large number of ways, and all have to be tried. (The example uses [!?] rather
1711: than a single character at the end, because both PCRE and Perl have an
1712: optimization that allows for fast failure when a single character is used. They
1713: remember the last single character that is required for a match, and fail early
1714: if it is not present in the string.) If the pattern is changed so that it uses
1715: an atomic group, like this:
1716: .sp
1717: ((?>\eD+)|<\ed+>)*[!?]
1718: .sp
1719: sequences of non-digits cannot be broken, and failure happens quickly.
1720: .
1721: .
1722: .\" HTML <a name="backreferences"></a>
1723: .SH "BACK REFERENCES"
1724: .rs
1725: .sp
1726: Outside a character class, a backslash followed by a digit greater than 0 (and
1727: possibly further digits) is a back reference to a capturing subpattern earlier
1728: (that is, to its left) in the pattern, provided there have been that many
1729: previous capturing left parentheses.
1730: .P
1731: However, if the decimal number following the backslash is less than 10, it is
1732: always taken as a back reference, and causes an error only if there are not
1733: that many capturing left parentheses in the entire pattern. In other words, the
1734: parentheses that are referenced need not be to the left of the reference for
1735: numbers less than 10. A "forward back reference" of this type can make sense
1736: when a repetition is involved and the subpattern to the right has participated
1737: in an earlier iteration.
1738: .P
1739: It is not possible to have a numerical "forward back reference" to a subpattern
1740: whose number is 10 or more using this syntax because a sequence such as \e50 is
1741: interpreted as a character defined in octal. See the subsection entitled
1742: "Non-printing characters"
1743: .\" HTML <a href="#digitsafterbackslash">
1744: .\" </a>
1745: above
1746: .\"
1747: for further details of the handling of digits following a backslash. There is
1748: no such problem when named parentheses are used. A back reference to any
1749: subpattern is possible using named parentheses (see below).
1750: .P
1751: Another way of avoiding the ambiguity inherent in the use of digits following a
1752: backslash is to use the \eg escape sequence. This escape must be followed by an
1753: unsigned number or a negative number, optionally enclosed in braces. These
1754: examples are all identical:
1755: .sp
1756: (ring), \e1
1757: (ring), \eg1
1758: (ring), \eg{1}
1759: .sp
1760: An unsigned number specifies an absolute reference without the ambiguity that
1761: is present in the older syntax. It is also useful when literal digits follow
1762: the reference. A negative number is a relative reference. Consider this
1763: example:
1764: .sp
1765: (abc(def)ghi)\eg{-1}
1766: .sp
1767: The sequence \eg{-1} is a reference to the most recently started capturing
1768: subpattern before \eg, that is, is it equivalent to \e2 in this example.
1769: Similarly, \eg{-2} would be equivalent to \e1. The use of relative references
1770: can be helpful in long patterns, and also in patterns that are created by
1771: joining together fragments that contain references within themselves.
1772: .P
1773: A back reference matches whatever actually matched the capturing subpattern in
1774: the current subject string, rather than anything matching the subpattern
1775: itself (see
1776: .\" HTML <a href="#subpatternsassubroutines">
1777: .\" </a>
1778: "Subpatterns as subroutines"
1779: .\"
1780: below for a way of doing that). So the pattern
1781: .sp
1782: (sens|respons)e and \e1ibility
1783: .sp
1784: matches "sense and sensibility" and "response and responsibility", but not
1785: "sense and responsibility". If caseful matching is in force at the time of the
1786: back reference, the case of letters is relevant. For example,
1787: .sp
1788: ((?i)rah)\es+\e1
1789: .sp
1790: matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1791: capturing subpattern is matched caselessly.
1792: .P
1793: There are several different ways of writing back references to named
1794: subpatterns. The .NET syntax \ek{name} and the Perl syntax \ek<name> or
1795: \ek'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
1796: back reference syntax, in which \eg can be used for both numeric and named
1797: references, is also supported. We could rewrite the above example in any of
1798: the following ways:
1799: .sp
1800: (?<p1>(?i)rah)\es+\ek<p1>
1801: (?'p1'(?i)rah)\es+\ek{p1}
1802: (?P<p1>(?i)rah)\es+(?P=p1)
1803: (?<p1>(?i)rah)\es+\eg{p1}
1804: .sp
1805: A subpattern that is referenced by name may appear in the pattern before or
1806: after the reference.
1807: .P
1808: There may be more than one back reference to the same subpattern. If a
1809: subpattern has not actually been used in a particular match, any back
1810: references to it always fail by default. For example, the pattern
1811: .sp
1812: (a|(bc))\e2
1813: .sp
1814: always fails if it starts to match "a" rather than "bc". However, if the
1815: PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back reference to an
1816: unset value matches an empty string.
1817: .P
1818: Because there may be many capturing parentheses in a pattern, all digits
1819: following a backslash are taken as part of a potential back reference number.
1820: If the pattern continues with a digit character, some delimiter must be used to
1821: terminate the back reference. If the PCRE_EXTENDED option is set, this can be
1822: whitespace. Otherwise, the \eg{ syntax or an empty comment (see
1823: .\" HTML <a href="#comments">
1824: .\" </a>
1825: "Comments"
1826: .\"
1827: below) can be used.
1828: .
1829: .SS "Recursive back references"
1830: .rs
1831: .sp
1832: A back reference that occurs inside the parentheses to which it refers fails
1833: when the subpattern is first used, so, for example, (a\e1) never matches.
1834: However, such references can be useful inside repeated subpatterns. For
1835: example, the pattern
1836: .sp
1837: (a|b\e1)+
1838: .sp
1839: matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1840: the subpattern, the back reference matches the character string corresponding
1841: to the previous iteration. In order for this to work, the pattern must be such
1842: that the first iteration does not need to match the back reference. This can be
1843: done using alternation, as in the example above, or by a quantifier with a
1844: minimum of zero.
1845: .P
1846: Back references of this type cause the group that they reference to be treated
1847: as an
1848: .\" HTML <a href="#atomicgroup">
1849: .\" </a>
1850: atomic group.
1851: .\"
1852: Once the whole group has been matched, a subsequent matching failure cannot
1853: cause backtracking into the middle of the group.
1854: .
1855: .
1856: .\" HTML <a name="bigassertions"></a>
1857: .SH ASSERTIONS
1858: .rs
1859: .sp
1860: An assertion is a test on the characters following or preceding the current
1861: matching point that does not actually consume any characters. The simple
1862: assertions coded as \eb, \eB, \eA, \eG, \eZ, \ez, ^ and $ are described
1863: .\" HTML <a href="#smallassertions">
1864: .\" </a>
1865: above.
1866: .\"
1867: .P
1868: More complicated assertions are coded as subpatterns. There are two kinds:
1869: those that look ahead of the current position in the subject string, and those
1870: that look behind it. An assertion subpattern is matched in the normal way,
1871: except that it does not cause the current matching position to be changed.
1872: .P
1873: Assertion subpatterns are not capturing subpatterns. If such an assertion
1874: contains capturing subpatterns within it, these are counted for the purposes of
1875: numbering the capturing subpatterns in the whole pattern. However, substring
1876: capturing is carried out only for positive assertions, because it does not make
1877: sense for negative assertions.
1878: .P
1879: For compatibility with Perl, assertion subpatterns may be repeated; though
1880: it makes no sense to assert the same thing several times, the side effect of
1881: capturing parentheses may occasionally be useful. In practice, there only three
1882: cases:
1883: .sp
1884: (1) If the quantifier is {0}, the assertion is never obeyed during matching.
1885: However, it may contain internal capturing parenthesized groups that are called
1886: from elsewhere via the
1887: .\" HTML <a href="#subpatternsassubroutines">
1888: .\" </a>
1889: subroutine mechanism.
1890: .\"
1891: .sp
1892: (2) If quantifier is {0,n} where n is greater than zero, it is treated as if it
1893: were {0,1}. At run time, the rest of the pattern match is tried with and
1894: without the assertion, the order depending on the greediness of the quantifier.
1895: .sp
1896: (3) If the minimum repetition is greater than zero, the quantifier is ignored.
1897: The assertion is obeyed just once when encountered during matching.
1898: .
1899: .
1900: .SS "Lookahead assertions"
1901: .rs
1902: .sp
1903: Lookahead assertions start with (?= for positive assertions and (?! for
1904: negative assertions. For example,
1905: .sp
1906: \ew+(?=;)
1907: .sp
1908: matches a word followed by a semicolon, but does not include the semicolon in
1909: the match, and
1910: .sp
1911: foo(?!bar)
1912: .sp
1913: matches any occurrence of "foo" that is not followed by "bar". Note that the
1914: apparently similar pattern
1915: .sp
1916: (?!foo)bar
1917: .sp
1918: does not find an occurrence of "bar" that is preceded by something other than
1919: "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1920: (?!foo) is always true when the next three characters are "bar". A
1921: lookbehind assertion is needed to achieve the other effect.
1922: .P
1923: If you want to force a matching failure at some point in a pattern, the most
1924: convenient way to do it is with (?!) because an empty string always matches, so
1925: an assertion that requires there not to be an empty string must always fail.
1926: The backtracking control verb (*FAIL) or (*F) is a synonym for (?!).
1927: .
1928: .
1929: .\" HTML <a name="lookbehind"></a>
1930: .SS "Lookbehind assertions"
1931: .rs
1932: .sp
1933: Lookbehind assertions start with (?<= for positive assertions and (?<! for
1934: negative assertions. For example,
1935: .sp
1936: (?<!foo)bar
1937: .sp
1938: does find an occurrence of "bar" that is not preceded by "foo". The contents of
1939: a lookbehind assertion are restricted such that all the strings it matches must
1940: have a fixed length. However, if there are several top-level alternatives, they
1941: do not all have to have the same fixed length. Thus
1942: .sp
1943: (?<=bullock|donkey)
1944: .sp
1945: is permitted, but
1946: .sp
1947: (?<!dogs?|cats?)
1948: .sp
1949: causes an error at compile time. Branches that match different length strings
1950: are permitted only at the top level of a lookbehind assertion. This is an
1951: extension compared with Perl, which requires all branches to match the same
1952: length of string. An assertion such as
1953: .sp
1954: (?<=ab(c|de))
1955: .sp
1956: is not permitted, because its single top-level branch can match two different
1957: lengths, but it is acceptable to PCRE if rewritten to use two top-level
1958: branches:
1959: .sp
1960: (?<=abc|abde)
1961: .sp
1962: In some cases, the escape sequence \eK
1963: .\" HTML <a href="#resetmatchstart">
1964: .\" </a>
1965: (see above)
1966: .\"
1967: can be used instead of a lookbehind assertion to get round the fixed-length
1968: restriction.
1969: .P
1970: The implementation of lookbehind assertions is, for each alternative, to
1971: temporarily move the current position back by the fixed length and then try to
1972: match. If there are insufficient characters before the current position, the
1973: assertion fails.
1974: .P
1975: In UTF-8 mode, PCRE does not allow the \eC escape (which matches a single byte,
1976: even in UTF-8 mode) to appear in lookbehind assertions, because it makes it
1977: impossible to calculate the length of the lookbehind. The \eX and \eR escapes,
1978: which can match different numbers of bytes, are also not permitted.
1979: .P
1980: .\" HTML <a href="#subpatternsassubroutines">
1981: .\" </a>
1982: "Subroutine"
1983: .\"
1984: calls (see below) such as (?2) or (?&X) are permitted in lookbehinds, as long
1985: as the subpattern matches a fixed-length string.
1986: .\" HTML <a href="#recursion">
1987: .\" </a>
1988: Recursion,
1989: .\"
1990: however, is not supported.
1991: .P
1992: Possessive quantifiers can be used in conjunction with lookbehind assertions to
1993: specify efficient matching of fixed-length strings at the end of subject
1994: strings. Consider a simple pattern such as
1995: .sp
1996: abcd$
1997: .sp
1998: when applied to a long string that does not match. Because matching proceeds
1999: from left to right, PCRE will look for each "a" in the subject and then see if
2000: what follows matches the rest of the pattern. If the pattern is specified as
2001: .sp
2002: ^.*abcd$
2003: .sp
2004: the initial .* matches the entire string at first, but when this fails (because
2005: there is no following "a"), it backtracks to match all but the last character,
2006: then all but the last two characters, and so on. Once again the search for "a"
2007: covers the entire string, from right to left, so we are no better off. However,
2008: if the pattern is written as
2009: .sp
2010: ^.*+(?<=abcd)
2011: .sp
2012: there can be no backtracking for the .*+ item; it can match only the entire
2013: string. The subsequent lookbehind assertion does a single test on the last four
2014: characters. If it fails, the match fails immediately. For long strings, this
2015: approach makes a significant difference to the processing time.
2016: .
2017: .
2018: .SS "Using multiple assertions"
2019: .rs
2020: .sp
2021: Several assertions (of any sort) may occur in succession. For example,
2022: .sp
2023: (?<=\ed{3})(?<!999)foo
2024: .sp
2025: matches "foo" preceded by three digits that are not "999". Notice that each of
2026: the assertions is applied independently at the same point in the subject
2027: string. First there is a check that the previous three characters are all
2028: digits, and then there is a check that the same three characters are not "999".
2029: This pattern does \fInot\fP match "foo" preceded by six characters, the first
2030: of which are digits and the last three of which are not "999". For example, it
2031: doesn't match "123abcfoo". A pattern to do that is
2032: .sp
2033: (?<=\ed{3}...)(?<!999)foo
2034: .sp
2035: This time the first assertion looks at the preceding six characters, checking
2036: that the first three are digits, and then the second assertion checks that the
2037: preceding three characters are not "999".
2038: .P
2039: Assertions can be nested in any combination. For example,
2040: .sp
2041: (?<=(?<!foo)bar)baz
2042: .sp
2043: matches an occurrence of "baz" that is preceded by "bar" which in turn is not
2044: preceded by "foo", while
2045: .sp
2046: (?<=\ed{3}(?!999)...)foo
2047: .sp
2048: is another pattern that matches "foo" preceded by three digits and any three
2049: characters that are not "999".
2050: .
2051: .
2052: .\" HTML <a name="conditions"></a>
2053: .SH "CONDITIONAL SUBPATTERNS"
2054: .rs
2055: .sp
2056: It is possible to cause the matching process to obey a subpattern
2057: conditionally or to choose between two alternative subpatterns, depending on
2058: the result of an assertion, or whether a specific capturing subpattern has
2059: already been matched. The two possible forms of conditional subpattern are:
2060: .sp
2061: (?(condition)yes-pattern)
2062: (?(condition)yes-pattern|no-pattern)
2063: .sp
2064: If the condition is satisfied, the yes-pattern is used; otherwise the
2065: no-pattern (if present) is used. If there are more than two alternatives in the
2066: subpattern, a compile-time error occurs. Each of the two alternatives may
2067: itself contain nested subpatterns of any form, including conditional
2068: subpatterns; the restriction to two alternatives applies only at the level of
2069: the condition. This pattern fragment is an example where the alternatives are
2070: complex:
2071: .sp
2072: (?(1) (A|B|C) | (D | (?(2)E|F) | E) )
2073: .sp
2074: .P
2075: There are four kinds of condition: references to subpatterns, references to
2076: recursion, a pseudo-condition called DEFINE, and assertions.
2077: .
2078: .SS "Checking for a used subpattern by number"
2079: .rs
2080: .sp
2081: If the text between the parentheses consists of a sequence of digits, the
2082: condition is true if a capturing subpattern of that number has previously
2083: matched. If there is more than one capturing subpattern with the same number
2084: (see the earlier
2085: .\"
2086: .\" HTML <a href="#recursion">
2087: .\" </a>
2088: section about duplicate subpattern numbers),
2089: .\"
2090: the condition is true if any of them have matched. An alternative notation is
2091: to precede the digits with a plus or minus sign. In this case, the subpattern
2092: number is relative rather than absolute. The most recently opened parentheses
2093: can be referenced by (?(-1), the next most recent by (?(-2), and so on. Inside
2094: loops it can also make sense to refer to subsequent groups. The next
2095: parentheses to be opened can be referenced as (?(+1), and so on. (The value
2096: zero in any of these forms is not used; it provokes a compile-time error.)
2097: .P
2098: Consider the following pattern, which contains non-significant white space to
2099: make it more readable (assume the PCRE_EXTENDED option) and to divide it into
2100: three parts for ease of discussion:
2101: .sp
2102: ( \e( )? [^()]+ (?(1) \e) )
2103: .sp
2104: The first part matches an optional opening parenthesis, and if that
2105: character is present, sets it as the first captured substring. The second part
2106: matches one or more characters that are not parentheses. The third part is a
2107: conditional subpattern that tests whether or not the first set of parentheses
2108: matched. If they did, that is, if subject started with an opening parenthesis,
2109: the condition is true, and so the yes-pattern is executed and a closing
2110: parenthesis is required. Otherwise, since no-pattern is not present, the
2111: subpattern matches nothing. In other words, this pattern matches a sequence of
2112: non-parentheses, optionally enclosed in parentheses.
2113: .P
2114: If you were embedding this pattern in a larger one, you could use a relative
2115: reference:
2116: .sp
2117: ...other stuff... ( \e( )? [^()]+ (?(-1) \e) ) ...
2118: .sp
2119: This makes the fragment independent of the parentheses in the larger pattern.
2120: .
2121: .SS "Checking for a used subpattern by name"
2122: .rs
2123: .sp
2124: Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a used
2125: subpattern by name. For compatibility with earlier versions of PCRE, which had
2126: this facility before Perl, the syntax (?(name)...) is also recognized. However,
2127: there is a possible ambiguity with this syntax, because subpattern names may
2128: consist entirely of digits. PCRE looks first for a named subpattern; if it
2129: cannot find one and the name consists entirely of digits, PCRE looks for a
2130: subpattern of that number, which must be greater than zero. Using subpattern
2131: names that consist entirely of digits is not recommended.
2132: .P
2133: Rewriting the above example to use a named subpattern gives this:
2134: .sp
2135: (?<OPEN> \e( )? [^()]+ (?(<OPEN>) \e) )
2136: .sp
2137: If the name used in a condition of this kind is a duplicate, the test is
2138: applied to all subpatterns of the same name, and is true if any one of them has
2139: matched.
2140: .
2141: .SS "Checking for pattern recursion"
2142: .rs
2143: .sp
2144: If the condition is the string (R), and there is no subpattern with the name R,
2145: the condition is true if a recursive call to the whole pattern or any
2146: subpattern has been made. If digits or a name preceded by ampersand follow the
2147: letter R, for example:
2148: .sp
2149: (?(R3)...) or (?(R&name)...)
2150: .sp
2151: the condition is true if the most recent recursion is into a subpattern whose
2152: number or name is given. This condition does not check the entire recursion
2153: stack. If the name used in a condition of this kind is a duplicate, the test is
2154: applied to all subpatterns of the same name, and is true if any one of them is
2155: the most recent recursion.
2156: .P
2157: At "top level", all these recursion test conditions are false.
2158: .\" HTML <a href="#recursion">
2159: .\" </a>
2160: The syntax for recursive patterns
2161: .\"
2162: is described below.
2163: .
2164: .\" HTML <a name="subdefine"></a>
2165: .SS "Defining subpatterns for use by reference only"
2166: .rs
2167: .sp
2168: If the condition is the string (DEFINE), and there is no subpattern with the
2169: name DEFINE, the condition is always false. In this case, there may be only one
2170: alternative in the subpattern. It is always skipped if control reaches this
2171: point in the pattern; the idea of DEFINE is that it can be used to define
2172: subroutines that can be referenced from elsewhere. (The use of
2173: .\" HTML <a href="#subpatternsassubroutines">
2174: .\" </a>
2175: subroutines
2176: .\"
2177: is described below.) For example, a pattern to match an IPv4 address such as
2178: "192.168.23.245" could be written like this (ignore whitespace and line
2179: breaks):
2180: .sp
2181: (?(DEFINE) (?<byte> 2[0-4]\ed | 25[0-5] | 1\ed\ed | [1-9]?\ed) )
2182: \eb (?&byte) (\e.(?&byte)){3} \eb
2183: .sp
2184: The first part of the pattern is a DEFINE group inside which a another group
2185: named "byte" is defined. This matches an individual component of an IPv4
2186: address (a number less than 256). When matching takes place, this part of the
2187: pattern is skipped because DEFINE acts like a false condition. The rest of the
2188: pattern uses references to the named group to match the four dot-separated
2189: components of an IPv4 address, insisting on a word boundary at each end.
2190: .
2191: .SS "Assertion conditions"
2192: .rs
2193: .sp
2194: If the condition is not in any of the above formats, it must be an assertion.
2195: This may be a positive or negative lookahead or lookbehind assertion. Consider
2196: this pattern, again containing non-significant white space, and with the two
2197: alternatives on the second line:
2198: .sp
2199: (?(?=[^a-z]*[a-z])
2200: \ed{2}-[a-z]{3}-\ed{2} | \ed{2}-\ed{2}-\ed{2} )
2201: .sp
2202: The condition is a positive lookahead assertion that matches an optional
2203: sequence of non-letters followed by a letter. In other words, it tests for the
2204: presence of at least one letter in the subject. If a letter is found, the
2205: subject is matched against the first alternative; otherwise it is matched
2206: against the second. This pattern matches strings in one of the two forms
2207: dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
2208: .
2209: .
2210: .\" HTML <a name="comments"></a>
2211: .SH COMMENTS
2212: .rs
2213: .sp
2214: There are two ways of including comments in patterns that are processed by
2215: PCRE. In both cases, the start of the comment must not be in a character class,
2216: nor in the middle of any other sequence of related characters such as (?: or a
2217: subpattern name or number. The characters that make up a comment play no part
2218: in the pattern matching.
2219: .P
2220: The sequence (?# marks the start of a comment that continues up to the next
2221: closing parenthesis. Nested parentheses are not permitted. If the PCRE_EXTENDED
2222: option is set, an unescaped # character also introduces a comment, which in
2223: this case continues to immediately after the next newline character or
2224: character sequence in the pattern. Which characters are interpreted as newlines
2225: is controlled by the options passed to \fBpcre_compile()\fP or by a special
2226: sequence at the start of the pattern, as described in the section entitled
2227: .\" HTML <a href="#newlines">
2228: .\" </a>
2229: "Newline conventions"
2230: .\"
2231: above. Note that the end of this type of comment is a literal newline sequence
2232: in the pattern; escape sequences that happen to represent a newline do not
2233: count. For example, consider this pattern when PCRE_EXTENDED is set, and the
2234: default newline convention is in force:
2235: .sp
2236: abc #comment \en still comment
2237: .sp
2238: On encountering the # character, \fBpcre_compile()\fP skips along, looking for
2239: a newline in the pattern. The sequence \en is still literal at this stage, so
2240: it does not terminate the comment. Only an actual character with the code value
2241: 0x0a (the default newline) does so.
2242: .
2243: .
2244: .\" HTML <a name="recursion"></a>
2245: .SH "RECURSIVE PATTERNS"
2246: .rs
2247: .sp
2248: Consider the problem of matching a string in parentheses, allowing for
2249: unlimited nested parentheses. Without the use of recursion, the best that can
2250: be done is to use a pattern that matches up to some fixed depth of nesting. It
2251: is not possible to handle an arbitrary nesting depth.
2252: .P
2253: For some time, Perl has provided a facility that allows regular expressions to
2254: recurse (amongst other things). It does this by interpolating Perl code in the
2255: expression at run time, and the code can refer to the expression itself. A Perl
2256: pattern using code interpolation to solve the parentheses problem can be
2257: created like this:
2258: .sp
2259: $re = qr{\e( (?: (?>[^()]+) | (?p{$re}) )* \e)}x;
2260: .sp
2261: The (?p{...}) item interpolates Perl code at run time, and in this case refers
2262: recursively to the pattern in which it appears.
2263: .P
2264: Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
2265: supports special syntax for recursion of the entire pattern, and also for
2266: individual subpattern recursion. After its introduction in PCRE and Python,
2267: this kind of recursion was subsequently introduced into Perl at release 5.10.
2268: .P
2269: A special item that consists of (? followed by a number greater than zero and a
2270: closing parenthesis is a recursive subroutine call of the subpattern of the
2271: given number, provided that it occurs inside that subpattern. (If not, it is a
2272: .\" HTML <a href="#subpatternsassubroutines">
2273: .\" </a>
2274: non-recursive subroutine
2275: .\"
2276: call, which is described in the next section.) The special item (?R) or (?0) is
2277: a recursive call of the entire regular expression.
2278: .P
2279: This PCRE pattern solves the nested parentheses problem (assume the
2280: PCRE_EXTENDED option is set so that white space is ignored):
2281: .sp
2282: \e( ( [^()]++ | (?R) )* \e)
2283: .sp
2284: First it matches an opening parenthesis. Then it matches any number of
2285: substrings which can either be a sequence of non-parentheses, or a recursive
2286: match of the pattern itself (that is, a correctly parenthesized substring).
2287: Finally there is a closing parenthesis. Note the use of a possessive quantifier
2288: to avoid backtracking into sequences of non-parentheses.
2289: .P
2290: If this were part of a larger pattern, you would not want to recurse the entire
2291: pattern, so instead you could use this:
2292: .sp
2293: ( \e( ( [^()]++ | (?1) )* \e) )
2294: .sp
2295: We have put the pattern into parentheses, and caused the recursion to refer to
2296: them instead of the whole pattern.
2297: .P
2298: In a larger pattern, keeping track of parenthesis numbers can be tricky. This
2299: is made easier by the use of relative references. Instead of (?1) in the
2300: pattern above you can write (?-2) to refer to the second most recently opened
2301: parentheses preceding the recursion. In other words, a negative number counts
2302: capturing parentheses leftwards from the point at which it is encountered.
2303: .P
2304: It is also possible to refer to subsequently opened parentheses, by writing
2305: references such as (?+2). However, these cannot be recursive because the
2306: reference is not inside the parentheses that are referenced. They are always
2307: .\" HTML <a href="#subpatternsassubroutines">
2308: .\" </a>
2309: non-recursive subroutine
2310: .\"
2311: calls, as described in the next section.
2312: .P
2313: An alternative approach is to use named parentheses instead. The Perl syntax
2314: for this is (?&name); PCRE's earlier syntax (?P>name) is also supported. We
2315: could rewrite the above example as follows:
2316: .sp
2317: (?<pn> \e( ( [^()]++ | (?&pn) )* \e) )
2318: .sp
2319: If there is more than one subpattern with the same name, the earliest one is
2320: used.
2321: .P
2322: This particular example pattern that we have been looking at contains nested
2323: unlimited repeats, and so the use of a possessive quantifier for matching
2324: strings of non-parentheses is important when applying the pattern to strings
2325: that do not match. For example, when this pattern is applied to
2326: .sp
2327: (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
2328: .sp
2329: it yields "no match" quickly. However, if a possessive quantifier is not used,
2330: the match runs for a very long time indeed because there are so many different
2331: ways the + and * repeats can carve up the subject, and all have to be tested
2332: before failure can be reported.
2333: .P
2334: At the end of a match, the values of capturing parentheses are those from
2335: the outermost level. If you want to obtain intermediate values, a callout
2336: function can be used (see below and the
2337: .\" HREF
2338: \fBpcrecallout\fP
2339: .\"
2340: documentation). If the pattern above is matched against
2341: .sp
2342: (ab(cd)ef)
2343: .sp
2344: the value for the inner capturing parentheses (numbered 2) is "ef", which is
2345: the last value taken on at the top level. If a capturing subpattern is not
2346: matched at the top level, its final captured value is unset, even if it was
2347: (temporarily) set at a deeper level during the matching process.
2348: .P
2349: If there are more than 15 capturing parentheses in a pattern, PCRE has to
2350: obtain extra memory to store data during a recursion, which it does by using
2351: \fBpcre_malloc\fP, freeing it via \fBpcre_free\fP afterwards. If no memory can
2352: be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
2353: .P
2354: Do not confuse the (?R) item with the condition (R), which tests for recursion.
2355: Consider this pattern, which matches text in angle brackets, allowing for
2356: arbitrary nesting. Only digits are allowed in nested brackets (that is, when
2357: recursing), whereas any characters are permitted at the outer level.
2358: .sp
2359: < (?: (?(R) \ed++ | [^<>]*+) | (?R)) * >
2360: .sp
2361: In this pattern, (?(R) is the start of a conditional subpattern, with two
2362: different alternatives for the recursive and non-recursive cases. The (?R) item
2363: is the actual recursive call.
2364: .
2365: .
2366: .\" HTML <a name="recursiondifference"></a>
2367: .SS "Differences in recursion processing between PCRE and Perl"
2368: .rs
2369: .sp
2370: Recursion processing in PCRE differs from Perl in two important ways. In PCRE
2371: (like Python, but unlike Perl), a recursive subpattern call is always treated
2372: as an atomic group. That is, once it has matched some of the subject string, it
2373: is never re-entered, even if it contains untried alternatives and there is a
2374: subsequent matching failure. This can be illustrated by the following pattern,
2375: which purports to match a palindromic string that contains an odd number of
2376: characters (for example, "a", "aba", "abcba", "abcdcba"):
2377: .sp
2378: ^(.|(.)(?1)\e2)$
2379: .sp
2380: The idea is that it either matches a single character, or two identical
2381: characters surrounding a sub-palindrome. In Perl, this pattern works; in PCRE
2382: it does not if the pattern is longer than three characters. Consider the
2383: subject string "abcba":
2384: .P
2385: At the top level, the first character is matched, but as it is not at the end
2386: of the string, the first alternative fails; the second alternative is taken
2387: and the recursion kicks in. The recursive call to subpattern 1 successfully
2388: matches the next character ("b"). (Note that the beginning and end of line
2389: tests are not part of the recursion).
2390: .P
2391: Back at the top level, the next character ("c") is compared with what
2392: subpattern 2 matched, which was "a". This fails. Because the recursion is
2393: treated as an atomic group, there are now no backtracking points, and so the
2394: entire match fails. (Perl is able, at this point, to re-enter the recursion and
2395: try the second alternative.) However, if the pattern is written with the
2396: alternatives in the other order, things are different:
2397: .sp
2398: ^((.)(?1)\e2|.)$
2399: .sp
2400: This time, the recursing alternative is tried first, and continues to recurse
2401: until it runs out of characters, at which point the recursion fails. But this
2402: time we do have another alternative to try at the higher level. That is the big
2403: difference: in the previous case the remaining alternative is at a deeper
2404: recursion level, which PCRE cannot use.
2405: .P
2406: To change the pattern so that it matches all palindromic strings, not just
2407: those with an odd number of characters, it is tempting to change the pattern to
2408: this:
2409: .sp
2410: ^((.)(?1)\e2|.?)$
2411: .sp
2412: Again, this works in Perl, but not in PCRE, and for the same reason. When a
2413: deeper recursion has matched a single character, it cannot be entered again in
2414: order to match an empty string. The solution is to separate the two cases, and
2415: write out the odd and even cases as alternatives at the higher level:
2416: .sp
2417: ^(?:((.)(?1)\e2|)|((.)(?3)\e4|.))
2418: .sp
2419: If you want to match typical palindromic phrases, the pattern has to ignore all
2420: non-word characters, which can be done like this:
2421: .sp
2422: ^\eW*+(?:((.)\eW*+(?1)\eW*+\e2|)|((.)\eW*+(?3)\eW*+\e4|\eW*+.\eW*+))\eW*+$
2423: .sp
2424: If run with the PCRE_CASELESS option, this pattern matches phrases such as "A
2425: man, a plan, a canal: Panama!" and it works well in both PCRE and Perl. Note
2426: the use of the possessive quantifier *+ to avoid backtracking into sequences of
2427: non-word characters. Without this, PCRE takes a great deal longer (ten times or
2428: more) to match typical phrases, and Perl takes so long that you think it has
2429: gone into a loop.
2430: .P
2431: \fBWARNING\fP: The palindrome-matching patterns above work only if the subject
2432: string does not start with a palindrome that is shorter than the entire string.
2433: For example, although "abcba" is correctly matched, if the subject is "ababa",
2434: PCRE finds the palindrome "aba" at the start, then fails at top level because
2435: the end of the string does not follow. Once again, it cannot jump back into the
2436: recursion to try other alternatives, so the entire match fails.
2437: .P
2438: The second way in which PCRE and Perl differ in their recursion processing is
2439: in the handling of captured values. In Perl, when a subpattern is called
2440: recursively or as a subpattern (see the next section), it has no access to any
2441: values that were captured outside the recursion, whereas in PCRE these values
2442: can be referenced. Consider this pattern:
2443: .sp
2444: ^(.)(\e1|a(?2))
2445: .sp
2446: In PCRE, this pattern matches "bab". The first capturing parentheses match "b",
2447: then in the second group, when the back reference \e1 fails to match "b", the
2448: second alternative matches "a" and then recurses. In the recursion, \e1 does
2449: now match "b" and so the whole match succeeds. In Perl, the pattern fails to
2450: match because inside the recursive call \e1 cannot access the externally set
2451: value.
2452: .
2453: .
2454: .\" HTML <a name="subpatternsassubroutines"></a>
2455: .SH "SUBPATTERNS AS SUBROUTINES"
2456: .rs
2457: .sp
2458: If the syntax for a recursive subpattern call (either by number or by
2459: name) is used outside the parentheses to which it refers, it operates like a
2460: subroutine in a programming language. The called subpattern may be defined
2461: before or after the reference. A numbered reference can be absolute or
2462: relative, as in these examples:
2463: .sp
2464: (...(absolute)...)...(?2)...
2465: (...(relative)...)...(?-1)...
2466: (...(?+1)...(relative)...
2467: .sp
2468: An earlier example pointed out that the pattern
2469: .sp
2470: (sens|respons)e and \e1ibility
2471: .sp
2472: matches "sense and sensibility" and "response and responsibility", but not
2473: "sense and responsibility". If instead the pattern
2474: .sp
2475: (sens|respons)e and (?1)ibility
2476: .sp
2477: is used, it does match "sense and responsibility" as well as the other two
2478: strings. Another example is given in the discussion of DEFINE above.
2479: .P
2480: All subroutine calls, whether recursive or not, are always treated as atomic
2481: groups. That is, once a subroutine has matched some of the subject string, it
2482: is never re-entered, even if it contains untried alternatives and there is a
2483: subsequent matching failure. Any capturing parentheses that are set during the
2484: subroutine call revert to their previous values afterwards.
2485: .P
2486: Processing options such as case-independence are fixed when a subpattern is
2487: defined, so if it is used as a subroutine, such options cannot be changed for
2488: different calls. For example, consider this pattern:
2489: .sp
2490: (abc)(?i:(?-1))
2491: .sp
2492: It matches "abcabc". It does not match "abcABC" because the change of
2493: processing option does not affect the called subpattern.
2494: .
2495: .
2496: .\" HTML <a name="onigurumasubroutines"></a>
2497: .SH "ONIGURUMA SUBROUTINE SYNTAX"
2498: .rs
2499: .sp
2500: For compatibility with Oniguruma, the non-Perl syntax \eg followed by a name or
2501: a number enclosed either in angle brackets or single quotes, is an alternative
2502: syntax for referencing a subpattern as a subroutine, possibly recursively. Here
2503: are two of the examples used above, rewritten using this syntax:
2504: .sp
2505: (?<pn> \e( ( (?>[^()]+) | \eg<pn> )* \e) )
2506: (sens|respons)e and \eg'1'ibility
2507: .sp
2508: PCRE supports an extension to Oniguruma: if a number is preceded by a
2509: plus or a minus sign it is taken as a relative reference. For example:
2510: .sp
2511: (abc)(?i:\eg<-1>)
2512: .sp
2513: Note that \eg{...} (Perl syntax) and \eg<...> (Oniguruma syntax) are \fInot\fP
2514: synonymous. The former is a back reference; the latter is a subroutine call.
2515: .
2516: .
2517: .SH CALLOUTS
2518: .rs
2519: .sp
2520: Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
2521: code to be obeyed in the middle of matching a regular expression. This makes it
2522: possible, amongst other things, to extract different substrings that match the
2523: same pair of parentheses when there is a repetition.
2524: .P
2525: PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
2526: code. The feature is called "callout". The caller of PCRE provides an external
2527: function by putting its entry point in the global variable \fIpcre_callout\fP.
2528: By default, this variable contains NULL, which disables all calling out.
2529: .P
2530: Within a regular expression, (?C) indicates the points at which the external
2531: function is to be called. If you want to identify different callout points, you
2532: can put a number less than 256 after the letter C. The default value is zero.
2533: For example, this pattern has two callout points:
2534: .sp
2535: (?C1)abc(?C2)def
2536: .sp
2537: If the PCRE_AUTO_CALLOUT flag is passed to \fBpcre_compile()\fP, callouts are
2538: automatically installed before each item in the pattern. They are all numbered
2539: 255.
2540: .P
2541: During matching, when PCRE reaches a callout point (and \fIpcre_callout\fP is
2542: set), the external function is called. It is provided with the number of the
2543: callout, the position in the pattern, and, optionally, one item of data
2544: originally supplied by the caller of \fBpcre_exec()\fP. The callout function
2545: may cause matching to proceed, to backtrack, or to fail altogether. A complete
2546: description of the interface to the callout function is given in the
2547: .\" HREF
2548: \fBpcrecallout\fP
2549: .\"
2550: documentation.
2551: .
2552: .
2553: .\" HTML <a name="backtrackcontrol"></a>
2554: .SH "BACKTRACKING CONTROL"
2555: .rs
2556: .sp
2557: Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
2558: are described in the Perl documentation as "experimental and subject to change
2559: or removal in a future version of Perl". It goes on to say: "Their usage in
2560: production code should be noted to avoid problems during upgrades." The same
2561: remarks apply to the PCRE features described in this section.
2562: .P
2563: Since these verbs are specifically related to backtracking, most of them can be
2564: used only when the pattern is to be matched using \fBpcre_exec()\fP, which uses
2565: a backtracking algorithm. With the exception of (*FAIL), which behaves like a
2566: failing negative assertion, they cause an error if encountered by
2567: \fBpcre_dfa_exec()\fP.
2568: .P
2569: If any of these verbs are used in an assertion or in a subpattern that is
2570: called as a subroutine (whether or not recursively), their effect is confined
2571: to that subpattern; it does not extend to the surrounding pattern, with one
2572: exception: the name from a *(MARK), (*PRUNE), or (*THEN) that is encountered in
2573: a successful positive assertion \fIis\fP passed back when a match succeeds
2574: (compare capturing parentheses in assertions). Note that such subpatterns are
2575: processed as anchored at the point where they are tested. Note also that Perl's
2576: treatment of subroutines is different in some cases.
2577: .P
2578: The new verbs make use of what was previously invalid syntax: an opening
2579: parenthesis followed by an asterisk. They are generally of the form
2580: (*VERB) or (*VERB:NAME). Some may take either form, with differing behaviour,
2581: depending on whether or not an argument is present. A name is any sequence of
2582: characters that does not include a closing parenthesis. If the name is empty,
2583: that is, if the closing parenthesis immediately follows the colon, the effect
2584: is as if the colon were not there. Any number of these verbs may occur in a
2585: pattern.
2586: .P
2587: PCRE contains some optimizations that are used to speed up matching by running
2588: some checks at the start of each match attempt. For example, it may know the
2589: minimum length of matching subject, or that a particular character must be
2590: present. When one of these optimizations suppresses the running of a match, any
2591: included backtracking verbs will not, of course, be processed. You can suppress
2592: the start-of-match optimizations by setting the PCRE_NO_START_OPTIMIZE option
2593: when calling \fBpcre_compile()\fP or \fBpcre_exec()\fP, or by starting the
2594: pattern with (*NO_START_OPT).
2595: .P
2596: Experiments with Perl suggest that it too has similar optimizations, sometimes
2597: leading to anomalous results.
2598: .
2599: .
2600: .SS "Verbs that act immediately"
2601: .rs
2602: .sp
2603: The following verbs act as soon as they are encountered. They may not be
2604: followed by a name.
2605: .sp
2606: (*ACCEPT)
2607: .sp
2608: This verb causes the match to end successfully, skipping the remainder of the
2609: pattern. However, when it is inside a subpattern that is called as a
2610: subroutine, only that subpattern is ended successfully. Matching then continues
2611: at the outer level. If (*ACCEPT) is inside capturing parentheses, the data so
2612: far is captured. For example:
2613: .sp
2614: A((?:A|B(*ACCEPT)|C)D)
2615: .sp
2616: This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is captured by
2617: the outer parentheses.
2618: .sp
2619: (*FAIL) or (*F)
2620: .sp
2621: This verb causes a matching failure, forcing backtracking to occur. It is
2622: equivalent to (?!) but easier to read. The Perl documentation notes that it is
2623: probably useful only when combined with (?{}) or (??{}). Those are, of course,
2624: Perl features that are not present in PCRE. The nearest equivalent is the
2625: callout feature, as for example in this pattern:
2626: .sp
2627: a+(?C)(*FAIL)
2628: .sp
2629: A match with the string "aaaa" always fails, but the callout is taken before
2630: each backtrack happens (in this example, 10 times).
2631: .
2632: .
2633: .SS "Recording which path was taken"
2634: .rs
2635: .sp
2636: There is one verb whose main purpose is to track how a match was arrived at,
2637: though it also has a secondary use in conjunction with advancing the match
2638: starting point (see (*SKIP) below).
2639: .sp
2640: (*MARK:NAME) or (*:NAME)
2641: .sp
2642: A name is always required with this verb. There may be as many instances of
2643: (*MARK) as you like in a pattern, and their names do not have to be unique.
2644: .P
2645: When a match succeeds, the name of the last-encountered (*MARK) on the matching
2646: path is passed back to the caller via the \fIpcre_extra\fP data structure, as
2647: described in the
2648: .\" HTML <a href="pcreapi.html#extradata">
2649: .\" </a>
2650: section on \fIpcre_extra\fP
2651: .\"
2652: in the
2653: .\" HREF
2654: \fBpcreapi\fP
2655: .\"
2656: documentation. Here is an example of \fBpcretest\fP output, where the /K
2657: modifier requests the retrieval and outputting of (*MARK) data:
2658: .sp
2659: re> /X(*MARK:A)Y|X(*MARK:B)Z/K
2660: data> XY
2661: 0: XY
2662: MK: A
2663: XZ
2664: 0: XZ
2665: MK: B
2666: .sp
2667: The (*MARK) name is tagged with "MK:" in this output, and in this example it
2668: indicates which of the two alternatives matched. This is a more efficient way
2669: of obtaining this information than putting each alternative in its own
2670: capturing parentheses.
2671: .P
2672: If (*MARK) is encountered in a positive assertion, its name is recorded and
2673: passed back if it is the last-encountered. This does not happen for negative
2674: assertions.
2675: .P
2676: After a partial match or a failed match, the name of the last encountered
2677: (*MARK) in the entire match process is returned. For example:
2678: .sp
2679: re> /X(*MARK:A)Y|X(*MARK:B)Z/K
2680: data> XP
2681: No match, mark = B
2682: .sp
2683: Note that in this unanchored example the mark is retained from the match
2684: attempt that started at the letter "X". Subsequent match attempts starting at
2685: "P" and then with an empty string do not get as far as the (*MARK) item, but
2686: nevertheless do not reset it.
2687: .
2688: .
2689: .SS "Verbs that act after backtracking"
2690: .rs
2691: .sp
2692: The following verbs do nothing when they are encountered. Matching continues
2693: with what follows, but if there is no subsequent match, causing a backtrack to
2694: the verb, a failure is forced. That is, backtracking cannot pass to the left of
2695: the verb. However, when one of these verbs appears inside an atomic group, its
2696: effect is confined to that group, because once the group has been matched,
2697: there is never any backtracking into it. In this situation, backtracking can
2698: "jump back" to the left of the entire atomic group. (Remember also, as stated
2699: above, that this localization also applies in subroutine calls and assertions.)
2700: .P
2701: These verbs differ in exactly what kind of failure occurs when backtracking
2702: reaches them.
2703: .sp
2704: (*COMMIT)
2705: .sp
2706: This verb, which may not be followed by a name, causes the whole match to fail
2707: outright if the rest of the pattern does not match. Even if the pattern is
2708: unanchored, no further attempts to find a match by advancing the starting point
2709: take place. Once (*COMMIT) has been passed, \fBpcre_exec()\fP is committed to
2710: finding a match at the current starting point, or not at all. For example:
2711: .sp
2712: a+(*COMMIT)b
2713: .sp
2714: This matches "xxaab" but not "aacaab". It can be thought of as a kind of
2715: dynamic anchor, or "I've started, so I must finish." The name of the most
2716: recently passed (*MARK) in the path is passed back when (*COMMIT) forces a
2717: match failure.
2718: .P
2719: Note that (*COMMIT) at the start of a pattern is not the same as an anchor,
2720: unless PCRE's start-of-match optimizations are turned off, as shown in this
2721: \fBpcretest\fP example:
2722: .sp
2723: re> /(*COMMIT)abc/
2724: data> xyzabc
2725: 0: abc
2726: xyzabc\eY
2727: No match
2728: .sp
2729: PCRE knows that any match must start with "a", so the optimization skips along
2730: the subject to "a" before running the first match attempt, which succeeds. When
2731: the optimization is disabled by the \eY escape in the second subject, the match
2732: starts at "x" and so the (*COMMIT) causes it to fail without trying any other
2733: starting points.
2734: .sp
2735: (*PRUNE) or (*PRUNE:NAME)
2736: .sp
2737: This verb causes the match to fail at the current starting position in the
2738: subject if the rest of the pattern does not match. If the pattern is
2739: unanchored, the normal "bumpalong" advance to the next starting character then
2740: happens. Backtracking can occur as usual to the left of (*PRUNE), before it is
2741: reached, or when matching to the right of (*PRUNE), but if there is no match to
2742: the right, backtracking cannot cross (*PRUNE). In simple cases, the use of
2743: (*PRUNE) is just an alternative to an atomic group or possessive quantifier,
2744: but there are some uses of (*PRUNE) that cannot be expressed in any other way.
2745: The behaviour of (*PRUNE:NAME) is the same as (*MARK:NAME)(*PRUNE). In an
2746: anchored pattern (*PRUNE) has the same effect as (*COMMIT).
2747: .sp
2748: (*SKIP)
2749: .sp
2750: This verb, when given without a name, is like (*PRUNE), except that if the
2751: pattern is unanchored, the "bumpalong" advance is not to the next character,
2752: but to the position in the subject where (*SKIP) was encountered. (*SKIP)
2753: signifies that whatever text was matched leading up to it cannot be part of a
2754: successful match. Consider:
2755: .sp
2756: a+(*SKIP)b
2757: .sp
2758: If the subject is "aaaac...", after the first match attempt fails (starting at
2759: the first character in the string), the starting point skips on to start the
2760: next attempt at "c". Note that a possessive quantifer does not have the same
2761: effect as this example; although it would suppress backtracking during the
2762: first match attempt, the second attempt would start at the second character
2763: instead of skipping on to "c".
2764: .sp
2765: (*SKIP:NAME)
2766: .sp
2767: When (*SKIP) has an associated name, its behaviour is modified. If the
2768: following pattern fails to match, the previous path through the pattern is
2769: searched for the most recent (*MARK) that has the same name. If one is found,
2770: the "bumpalong" advance is to the subject position that corresponds to that
2771: (*MARK) instead of to where (*SKIP) was encountered. If no (*MARK) with a
2772: matching name is found, the (*SKIP) is ignored.
2773: .sp
2774: (*THEN) or (*THEN:NAME)
2775: .sp
2776: This verb causes a skip to the next innermost alternative if the rest of the
2777: pattern does not match. That is, it cancels pending backtracking, but only
2778: within the current alternative. Its name comes from the observation that it can
2779: be used for a pattern-based if-then-else block:
2780: .sp
2781: ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
2782: .sp
2783: If the COND1 pattern matches, FOO is tried (and possibly further items after
2784: the end of the group if FOO succeeds); on failure, the matcher skips to the
2785: second alternative and tries COND2, without backtracking into COND1. The
2786: behaviour of (*THEN:NAME) is exactly the same as (*MARK:NAME)(*THEN).
2787: If (*THEN) is not inside an alternation, it acts like (*PRUNE).
2788: .P
2789: Note that a subpattern that does not contain a | character is just a part of
2790: the enclosing alternative; it is not a nested alternation with only one
2791: alternative. The effect of (*THEN) extends beyond such a subpattern to the
2792: enclosing alternative. Consider this pattern, where A, B, etc. are complex
2793: pattern fragments that do not contain any | characters at this level:
2794: .sp
2795: A (B(*THEN)C) | D
2796: .sp
2797: If A and B are matched, but there is a failure in C, matching does not
2798: backtrack into A; instead it moves to the next alternative, that is, D.
2799: However, if the subpattern containing (*THEN) is given an alternative, it
2800: behaves differently:
2801: .sp
2802: A (B(*THEN)C | (*FAIL)) | D
2803: .sp
2804: The effect of (*THEN) is now confined to the inner subpattern. After a failure
2805: in C, matching moves to (*FAIL), which causes the whole subpattern to fail
2806: because there are no more alternatives to try. In this case, matching does now
2807: backtrack into A.
2808: .P
2809: Note also that a conditional subpattern is not considered as having two
2810: alternatives, because only one is ever used. In other words, the | character in
2811: a conditional subpattern has a different meaning. Ignoring white space,
2812: consider:
2813: .sp
2814: ^.*? (?(?=a) a | b(*THEN)c )
2815: .sp
2816: If the subject is "ba", this pattern does not match. Because .*? is ungreedy,
2817: it initially matches zero characters. The condition (?=a) then fails, the
2818: character "b" is matched, but "c" is not. At this point, matching does not
2819: backtrack to .*? as might perhaps be expected from the presence of the |
2820: character. The conditional subpattern is part of the single alternative that
2821: comprises the whole pattern, and so the match fails. (If there was a backtrack
2822: into .*?, allowing it to match "b", the match would succeed.)
2823: .P
2824: The verbs just described provide four different "strengths" of control when
2825: subsequent matching fails. (*THEN) is the weakest, carrying on the match at the
2826: next alternative. (*PRUNE) comes next, failing the match at the current
2827: starting position, but allowing an advance to the next character (for an
2828: unanchored pattern). (*SKIP) is similar, except that the advance may be more
2829: than one character. (*COMMIT) is the strongest, causing the entire match to
2830: fail.
2831: .P
2832: If more than one such verb is present in a pattern, the "strongest" one wins.
2833: For example, consider this pattern, where A, B, etc. are complex pattern
2834: fragments:
2835: .sp
2836: (A(*COMMIT)B(*THEN)C|D)
2837: .sp
2838: Once A has matched, PCRE is committed to this match, at the current starting
2839: position. If subsequently B matches, but C does not, the normal (*THEN) action
2840: of trying the next alternative (that is, D) does not happen because (*COMMIT)
2841: overrides.
2842: .
2843: .
2844: .SH "SEE ALSO"
2845: .rs
2846: .sp
2847: \fBpcreapi\fP(3), \fBpcrecallout\fP(3), \fBpcrematching\fP(3),
2848: \fBpcresyntax\fP(3), \fBpcre\fP(3).
2849: .
2850: .
2851: .SH AUTHOR
2852: .rs
2853: .sp
2854: .nf
2855: Philip Hazel
2856: University Computing Service
2857: Cambridge CB2 3QH, England.
2858: .fi
2859: .
2860: .
2861: .SH REVISION
2862: .rs
2863: .sp
2864: .nf
2865: Last updated: 29 November 2011
2866: Copyright (c) 1997-2011 University of Cambridge.
2867: .fi
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>