/[pcre]/code/tags/pcre-2.02/pcre.3
ViewVC logotype

Contents of /code/tags/pcre-2.02/pcre.3

Parent Directory Parent Directory | Revision Log Revision Log


Revision 27 - (hide annotations) (download)
Sat Feb 24 21:38:49 2007 UTC (7 years, 6 months ago) by nigel
Original Path: code/trunk/pcre.3
File size: 58500 byte(s)
Load pcre-2.02 into code/trunk.

1 nigel 3 .TH PCRE 3
2     .SH NAME
3     pcre - Perl-compatible regular expressions.
4     .SH SYNOPSIS
5     .B #include <pcre.h>
6     .PP
7     .SM
8     .br
9     .B pcre *pcre_compile(const char *\fIpattern\fR, int \fIoptions\fR,
10     .ti +5n
11 nigel 25 .B const char **\fIerrptr\fR, int *\fIerroffset\fR,
12     .ti +5n
13     .B const unsigned char *\fItableptr\fR);
14 nigel 3 .PP
15     .br
16 nigel 25 .B const unsigned char *pcre_maketables(void);
17     .PP
18     .br
19 nigel 3 .B pcre_extra *pcre_study(const pcre *\fIcode\fR, int \fIoptions\fR,
20     .ti +5n
21 nigel 11 .B const char **\fIerrptr\fR);
22 nigel 3 .PP
23     .br
24     .B int pcre_exec(const pcre *\fIcode\fR, "const pcre_extra *\fIextra\fR,"
25     .ti +5n
26     .B "const char *\fIsubject\fR," int \fIlength\fR, int \fIoptions\fR,
27     .ti +5n
28     .B int *\fIovector\fR, int \fIovecsize\fR);
29     .PP
30     .br
31     .B int pcre_info(const pcre *\fIcode\fR, int *\fIoptptr\fR, int
32     .B *\fIfirstcharptr\fR);
33     .PP
34     .br
35     .B char *pcre_version(void);
36     .PP
37     .br
38     .B void *(*pcre_malloc)(size_t);
39     .PP
40     .br
41     .B void (*pcre_free)(void *);
42    
43    
44    
45     .SH DESCRIPTION
46     The PCRE library is a set of functions that implement regular expression
47     pattern matching using the same syntax and semantics as Perl 5, with just a few
48 nigel 23 differences (see below). The current implementation corresponds to Perl 5.005.
49 nigel 3
50     PCRE has its own native API, which is described in this man page. There is also
51     a set of wrapper functions that correspond to the POSIX API. See
52     \fBpcreposix (3)\fR.
53    
54     The three functions \fBpcre_compile()\fR, \fBpcre_study()\fR, and
55     \fBpcre_exec()\fR are used for compiling and matching regular expressions. The
56 nigel 25 function \fBpcre_maketables()\fR is used (optionally) to build a set of
57     character tables in the current locale for passing to \fBpcre_compile()\fR.
58    
59     The function \fBpcre_info()\fR is used to find out information about a compiled
60 nigel 3 pattern, while the function \fBpcre_version()\fR returns a pointer to a string
61     containing the version of PCRE and its date of release.
62    
63     The global variables \fBpcre_malloc\fR and \fBpcre_free\fR initially contain
64     the entry points of the standard \fBmalloc()\fR and \fBfree()\fR functions
65     respectively. PCRE calls the memory management functions via these variables,
66     so a calling program can replace them if it wishes to intercept the calls. This
67     should be done before calling any PCRE functions.
68    
69    
70     .SH MULTI-THREADING
71     The PCRE functions can be used in multi-threading applications, with the
72 nigel 25 proviso that the memory management functions pointed to by \fBpcre_malloc\fR
73     and \fBpcre_free\fR are shared by all threads.
74 nigel 3
75     The compiled form of a regular expression is not altered during matching, so
76     the same compiled pattern can safely be used by several threads at once.
77    
78    
79     .SH COMPILING A PATTERN
80     The function \fBpcre_compile()\fR is called to compile a pattern into an
81     internal form. The pattern is a C string terminated by a binary zero, and
82 nigel 25 is passed in the argument \fIpattern\fR. A pointer to a single block of memory
83     that is obtained via \fBpcre_malloc\fR is returned. This contains the
84     compiled code and related data. The \fBpcre\fR type is defined for this for
85     convenience, but in fact \fBpcre\fR is just a typedef for \fBvoid\fR, since the
86     contents of the block are not externally defined. It is up to the caller to
87     free the memory when it is no longer required.
88 nigel 3 .PP
89     The size of a compiled pattern is roughly proportional to the length of the
90     pattern string, except that each character class (other than those containing
91     just a single character, negated or not) requires 33 bytes, and repeat
92     quantifiers with a minimum greater than one or a bounded maximum cause the
93     relevant portions of the compiled pattern to be replicated.
94     .PP
95     The \fIoptions\fR argument contains independent bits that affect the
96 nigel 23 compilation. It should be zero if no options are required. Some of the options,
97     in particular, those that are compatible with Perl, can also be set and unset
98     from within the pattern (see the detailed description of regular expressions
99     below). For these options, the contents of the \fIoptions\fR argument specifies
100     their initial settings at the start of compilation and execution. The
101     PCRE_ANCHORED option can be set at the time of matching as well as at compile
102     time.
103 nigel 3 .PP
104     If \fIerrptr\fR is NULL, \fBpcre_compile()\fR returns NULL immediately.
105     Otherwise, if compilation of a pattern fails, \fBpcre_compile()\fR returns
106     NULL, and sets the variable pointed to by \fIerrptr\fR to point to a textual
107 nigel 25 error message. The offset from the start of the pattern to the character where
108     the error was discovered is placed in the variable pointed to by
109     \fIerroffset\fR, which must not be NULL. If it is, an immediate error is given.
110 nigel 3 .PP
111 nigel 25 If the final argument, \fItableptr\fR, is NULL, PCRE uses a default set of
112     character tables which are built when it is compiled, using the default C
113     locale. Otherwise, \fItableptr\fR must be the result of a call to
114     \fBpcre_maketables()\fR. See the section on locale support below.
115     .PP
116 nigel 3 The following option bits are defined in the header file:
117    
118     PCRE_ANCHORED
119    
120     If this bit is set, the pattern is forced to be "anchored", that is, it is
121     constrained to match only at the start of the string which is being searched
122     (the "subject string"). This effect can also be achieved by appropriate
123     constructs in the pattern itself, which is the only way to do it in Perl.
124    
125     PCRE_CASELESS
126    
127     If this bit is set, letters in the pattern match both upper and lower case
128 nigel 23 letters. It is equivalent to Perl's /i option.
129 nigel 3
130     PCRE_DOLLAR_ENDONLY
131    
132     If this bit is set, a dollar metacharacter in the pattern matches only at the
133 nigel 23 end of the subject string. Without this option, a dollar also matches
134     immediately before the final character if it is a newline (but not before any
135     other newlines). The PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is
136     set. There is no equivalent to this option in Perl.
137 nigel 3
138     PCRE_DOTALL
139    
140     If this bit is set, a dot metacharater in the pattern matches all characters,
141 nigel 23 including newlines. Without it, newlines are excluded. This option is
142 nigel 3 equivalent to Perl's /s option. A negative class such as [^a] always matches a
143     newline character, independent of the setting of this option.
144    
145     PCRE_EXTENDED
146    
147 nigel 23 If this bit is set, whitespace data characters in the pattern are totally
148     ignored except when escaped or inside a character class, and characters between
149     an unescaped # outside a character class and the next newline character,
150 nigel 3 inclusive, are also ignored. This is equivalent to Perl's /x option, and makes
151 nigel 23 it possible to include comments inside complicated patterns. Note, however,
152     that this applies only to data characters. Whitespace characters may never
153     appear within special character sequences in a pattern, for example within the
154     sequence (?( which introduces a conditional subpattern.
155 nigel 3
156 nigel 23 PCRE_EXTRA
157    
158     This option turns on additional functionality of PCRE that is incompatible with
159     Perl. Any backslash in a pattern that is followed by a letter that has no
160     special meaning causes an error, thus reserving these combinations for future
161     expansion. By default, as in Perl, a backslash followed by a letter with no
162     special meaning is treated as a literal. There are at present no other features
163     controlled by this option.
164    
165 nigel 3 PCRE_MULTILINE
166    
167     By default, PCRE treats the subject string as consisting of a single "line" of
168     characters (even if it actually contains several newlines). The "start of line"
169     metacharacter (^) matches only at the start of the string, while the "end of
170     line" metacharacter ($) matches only at the end of the string, or before a
171 nigel 23 terminating newline (unless PCRE_DOLLAR_ENDONLY is set). This is the same as
172     Perl.
173 nigel 3
174     When PCRE_MULTILINE it is set, the "start of line" and "end of line" constructs
175     match immediately following or immediately before any newline in the subject
176     string, respectively, as well as at the very start and end. This is equivalent
177     to Perl's /m option. If there are no "\\n" characters in a subject string, or
178     no occurrences of ^ or $ in a pattern, setting PCRE_MULTILINE has no
179     effect.
180    
181 nigel 19 PCRE_UNGREEDY
182 nigel 3
183 nigel 19 This option inverts the "greediness" of the quantifiers so that they are not
184     greedy by default, but become greedy if followed by "?". It is not compatible
185     with Perl. It can also be set by a (?U) option setting within the pattern.
186 nigel 3
187 nigel 19
188 nigel 3 .SH STUDYING A PATTERN
189     When a pattern is going to be used several times, it is worth spending more
190     time analyzing it in order to speed up the time taken for matching. The
191     function \fBpcre_study()\fR takes a pointer to a compiled pattern as its first
192     argument, and returns a pointer to a \fBpcre_extra\fR block (another \fBvoid\fR
193     typedef) containing additional information about the pattern; this can be
194     passed to \fBpcre_exec()\fR. If no additional information is available, NULL
195     is returned.
196    
197 nigel 23 The second argument contains option bits. At present, no options are defined
198     for \fBpcre_study()\fR, and this argument should always be zero.
199 nigel 3
200     The third argument for \fBpcre_study()\fR is a pointer to an error message. If
201     studying succeeds (even if no data is returned), the variable it points to is
202     set to NULL. Otherwise it points to a textual error message.
203    
204     At present, studying a pattern is useful only for non-anchored patterns that do
205     not have a single fixed starting character. A bitmap of possible starting
206     characters is created.
207    
208    
209 nigel 25 .SH LOCALE SUPPORT
210     PCRE handles caseless matching, and determines whether characters are letters,
211     digits, or whatever, by reference to a set of tables. The library contains a
212     default set of tables which is created in the default C locale when PCRE is
213     compiled. This is used when the final argument of \fBpcre_compile()\fR is NULL,
214     and is sufficient for many applications.
215    
216     An alternative set of tables can, however, be supplied. Such tables are built
217     by calling the \fBpcre_maketables()\fR function, which has no arguments, in the
218     relevant locale. The result can then be passed to \fBpcre_compile()\ as often
219     as necessary. For example, to build and use tables that are appropriate for the
220     French locale (where accented characters with codes greater than 128 are
221     treated as letters), the following code could be used:
222    
223     setlocale(LC_CTYPE, "fr");
224     tables = pcre_maketables();
225     re = pcre_compile(..., tables);
226    
227     The tables are built in memory that is obtained via \fBpcre_malloc\fR. The
228     pointer that is passed to \fBpcre_compile\fR is saved with the compiled
229     pattern, and the same tables are used via this pointer by \fBpcre_study()\fR
230     and \fBpcre_match()\fR. Thus for any single pattern, compilation, studying and
231     matching all happen in the same locale, but different patterns can be compiled
232     in different locales. It is the caller's responsibility to ensure that the
233     memory containing the tables remains available for as long as it is needed.
234    
235    
236 nigel 3 .SH MATCHING A PATTERN
237     The function \fBpcre_exec()\fR is called to match a subject string against a
238     pre-compiled pattern, which is passed in the \fIcode\fR argument. If the
239     pattern has been studied, the result of the study should be passed in the
240     \fIextra\fR argument. Otherwise this must be NULL.
241    
242     The subject string is passed as a pointer in \fIsubject\fR and a length in
243     \fIlength\fR. Unlike the pattern string, it may contain binary zero characters.
244    
245 nigel 23 The PCRE_ANCHORED option can be passed in the \fIoptions\fR argument, whose
246     unused bits must be zero. However, if a pattern was compiled with
247     PCRE_ANCHORED, or turned out to be anchored by virtue of its contents, it
248     cannot be made unachored at matching time.
249 nigel 3
250     There are also two further options that can be set only at matching time:
251    
252     PCRE_NOTBOL
253    
254     The first character of the string is not the beginning of a line, so the
255     circumflex metacharacter should not match before it. Setting this without
256 nigel 23 PCRE_MULTILINE (at compile time) causes circumflex never to match.
257 nigel 3
258     PCRE_NOTEOL
259    
260     The end of the string is not the end of a line, so the dollar metacharacter
261 nigel 23 should not match it nor (except in multiline mode) a newline immediately before
262     it. Setting this without PCRE_MULTILINE (at compile time) causes dollar never
263     to match.
264 nigel 3
265     In general, a pattern matches a certain portion of the subject, and in
266     addition, further substrings from the subject may be picked out by parts of the
267     pattern. Following the usage in Jeffrey Friedl's book, this is called
268     "capturing" in what follows, and the phrase "capturing subpattern" is used for
269     a fragment of a pattern that picks out a substring. PCRE supports several other
270     kinds of parenthesized subpattern that do not cause substrings to be captured.
271    
272     Captured substrings are returned to the caller via a vector of integer offsets
273     whose address is passed in \fIovector\fR. The number of elements in the vector
274 nigel 23 is passed in \fIovecsize\fR. The first two-thirds of the vector is used to pass
275     back captured substrings, each substring using a pair of integers. The
276     remaining third of the vector is used as workspace by \fBpcre_exec()\fR while
277     matching capturing subpatterns, and is not available for passing back
278     information. The length passed in \fIovecsize\fR should always be a multiple of
279     three. If it is not, it is rounded down.
280 nigel 3
281 nigel 23 When a match has been successful, information about captured substrings is
282     returned in pairs of integers, starting at the beginning of \fIovector\fR, and
283     continuing up to two-thirds of its length at the most. The first element of a
284     pair is set to the offset of the first character in a substring, and the second
285     is set to the offset of the first character after the end of a substring. The
286     first pair, \fIovector[0]\fR and \fIovector[1]\fR, identify the portion of the
287     subject string matched by the entire pattern. The next pair is used for the
288     first capturing subpattern, and so on. The value returned by \fBpcre_exec()\fR
289     is the number of pairs that have been set. If there are no capturing
290     subpatterns, the return value from a successful match is 1, indicating that
291     just the first pair of offsets has been set.
292 nigel 3
293     It is possible for an capturing subpattern number \fIn+1\fR to match some
294     part of the subject when subpattern \fIn\fR has not been used at all. For
295 nigel 23 example, if the string "abc" is matched against the pattern (a|(z))(bc)
296 nigel 3 subpatterns 1 and 3 are matched, but 2 is not. When this happens, both offset
297     values corresponding to the unused subpattern are set to -1.
298    
299     If a capturing subpattern is matched repeatedly, it is the last portion of the
300     string that it matched that gets returned.
301    
302     If the vector is too small to hold all the captured substrings, it is used as
303 nigel 23 far as possible (up to two-thirds of its length), and the function returns a
304     value of zero. In particular, if the substring offsets are not of interest,
305     \fBpcre_exec()\fR may be called with \fIovector\fR passed as NULL and
306     \fIovecsize\fR as zero. However, if the pattern contains back references and
307     the \fIovector\fR isn't big enough to remember the related substrings, PCRE has
308     to get additional memory for use during matching. Thus it is usually advisable
309     to supply an \fIovector\fR.
310 nigel 3
311     Note that \fBpcre_info()\fR can be used to find out how many capturing
312 nigel 23 subpatterns there are in a compiled pattern. The smallest size for
313     \fIovector\fR that will allow for \fIn\fR captured substrings in addition to
314     the offsets of the substring matched by the whole pattern is (\fIn\fR+1)*3.
315 nigel 3
316     If \fBpcre_exec()\fR fails, it returns a negative number. The following are
317     defined in the header file:
318    
319     PCRE_ERROR_NOMATCH (-1)
320    
321     The subject string did not match the pattern.
322    
323 nigel 23 PCRE_ERROR_NULL (-2)
324 nigel 3
325     Either \fIcode\fR or \fIsubject\fR was passed as NULL, or \fIovector\fR was
326     NULL and \fIovecsize\fR was not zero.
327    
328 nigel 23 PCRE_ERROR_BADOPTION (-3)
329 nigel 3
330     An unrecognized bit was set in the \fIoptions\fR argument.
331    
332 nigel 23 PCRE_ERROR_BADMAGIC (-4)
333 nigel 3
334     PCRE stores a 4-byte "magic number" at the start of the compiled code, to catch
335     the case when it is passed a junk pointer. This is the error it gives when the
336     magic number isn't present.
337    
338 nigel 23 PCRE_ERROR_UNKNOWN_NODE (-5)
339 nigel 3
340     While running the pattern match, an unknown item was encountered in the
341     compiled pattern. This error could be caused by a bug in PCRE or by overwriting
342     of the compiled pattern.
343    
344 nigel 23 PCRE_ERROR_NOMEMORY (-6)
345 nigel 3
346     If a pattern contains back references, but the \fIovector\fR that is passed to
347     \fBpcre_exec()\fR is not big enough to remember the referenced substrings, PCRE
348     gets a block of memory at the start of matching to use for this purpose. If the
349     call via \fBpcre_malloc()\fR fails, this error is given. The memory is freed at
350     the end of matching.
351    
352    
353     .SH INFORMATION ABOUT A PATTERN
354     The \fBpcre_info()\fR function returns information about a compiled pattern.
355     Its yield is the number of capturing subpatterns, or one of the following
356     negative numbers:
357    
358     PCRE_ERROR_NULL the argument \fIcode\fR was NULL
359     PCRE_ERROR_BADMAGIC the "magic number" was not found
360    
361     If the \fIoptptr\fR argument is not NULL, a copy of the options with which the
362     pattern was compiled is placed in the integer it points to.
363    
364     If the \fIfirstcharptr\fR argument is not NULL, is is used to pass back
365     information about the first character of any matched string. If there is a
366     fixed first character, e.g. from a pattern such as (cat|cow|coyote), then it is
367     returned in the integer pointed to by \fIfirstcharptr\fR. Otherwise, if the
368     pattern was compiled with the PCRE_MULTILINE option, and every branch started
369     with "^", then -1 is returned, indicating that the pattern will match at the
370     start of a subject string or after any "\\n" within the string. Otherwise -2 is
371     returned.
372    
373    
374     .SH LIMITATIONS
375     There are some size limitations in PCRE but it is hoped that they will never in
376     practice be relevant.
377     The maximum length of a compiled pattern is 65539 (sic) bytes.
378     All values in repeating quantifiers must be less than 65536.
379     The maximum number of capturing subpatterns is 99.
380     The maximum number of all parenthesized subpatterns, including capturing
381 nigel 23 subpatterns, assertions, and other types of subpattern, is 200.
382 nigel 3
383     The maximum length of a subject string is the largest positive number that an
384     integer variable can hold. However, PCRE uses recursion to handle subpatterns
385     and indefinite repetition. This means that the available stack space may limit
386     the size of a subject string that can be processed by certain patterns.
387    
388    
389     .SH DIFFERENCES FROM PERL
390 nigel 23 The differences described here are with respect to Perl 5.005.
391 nigel 3
392     1. By default, a whitespace character is any character that the C library
393     function \fBisspace()\fR recognizes, though it is possible to compile PCRE with
394     alternative character type tables. Normally \fBisspace()\fR matches space,
395     formfeed, newline, carriage return, horizontal tab, and vertical tab. Perl 5
396     no longer includes vertical tab in its set of whitespace characters. The \\v
397     escape that was in the Perl documentation for a long time was never in fact
398     recognized. However, the character itself was treated as whitespace at least
399 nigel 23 up to 5.002. In 5.004 and 5.005 it does not match \\s.
400 nigel 3
401     2. PCRE does not allow repeat quantifiers on lookahead assertions. Perl permits
402 nigel 23 them, but they do not mean what you might think. For example, (?!a){3} does
403 nigel 3 not assert that the next three characters are not "a". It just asserts that the
404     next character is not "a" three times.
405    
406     3. Capturing subpatterns that occur inside negative lookahead assertions are
407     counted, but their entries in the offsets vector are never set. Perl sets its
408     numerical variables from any such patterns that are matched before the
409     assertion fails to match something (thereby succeeding), but only if the
410     negative lookahead assertion contains just one branch.
411    
412     4. Though binary zero characters are supported in the subject string, they are
413     not allowed in a pattern string because it is passed as a normal C string,
414     terminated by zero. The escape sequence "\\0" can be used in the pattern to
415     represent a binary zero.
416    
417     5. The following Perl escape sequences are not supported: \\l, \\u, \\L, \\U,
418     \\E, \\Q. In fact these are implemented by Perl's general string-handling and
419     are not part of its pattern matching engine.
420    
421     6. The Perl \\G assertion is not supported as it is not relevant to single
422     pattern matches.
423    
424 nigel 23 7. Fairly obviously, PCRE does not support the (?{code}) construction.
425 nigel 3
426 nigel 23 8. There are at the time of writing some oddities in Perl 5.005_02 concerned
427     with the settings of captured strings when part of a pattern is repeated. For
428     example, matching "aba" against the pattern /^(a(b)?)+$/ sets $2 to the value
429     "b", but matching "aabbaa" against /^(aa(bb)?)+$/ leaves $2 unset. However, if
430     the pattern is changed to /^(aa(b(b))?)+$/ then $2 (and $3) get set.
431 nigel 3
432 nigel 23 In Perl 5.004 $2 is set in both cases, and that is also true of PCRE. If in the
433     future Perl changes to a consistent state that is different, PCRE may change to
434     follow.
435 nigel 3
436 nigel 23 9. Another as yet unresolved discrepancy is that in Perl 5.005_02 the pattern
437     /^(a)?(?(1)a|b)+$/ matches the string "a", whereas in PCRE it does not.
438     However, in both Perl and PCRE /^(a)?a/ matched against "a" leaves $1 unset.
439 nigel 3
440 nigel 23 10. PCRE provides some extensions to the Perl regular expression facilities:
441 nigel 3
442 nigel 23 (a) Although lookbehind assertions must match fixed length strings, each
443     alternative branch of a lookbehind assertion can match a different length of
444     string. Perl 5.005 requires them all to have the same length.
445 nigel 3
446 nigel 23 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $ meta-
447 nigel 3 character matches only at the very end of the string.
448    
449 nigel 23 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no special
450     meaning is faulted.
451 nigel 3
452 nigel 23 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quantifiers is
453 nigel 19 inverted, that is, by default they are not greedy, but if followed by a
454     question mark they are.
455 nigel 3
456 nigel 19
457 nigel 3 .SH REGULAR EXPRESSION DETAILS
458     The syntax and semantics of the regular expressions supported by PCRE are
459     described below. Regular expressions are also described in the Perl
460     documentation and in a number of other books, some of which have copious
461     examples. Jeffrey Friedl's "Mastering Regular Expressions", published by
462     O'Reilly (ISBN 1-56592-257-3), covers them in great detail. The description
463     here is intended as reference documentation.
464    
465     A regular expression is a pattern that is matched against a subject string from
466     left to right. Most characters stand for themselves in a pattern, and match the
467     corresponding characters in the subject. As a trivial example, the pattern
468    
469     The quick brown fox
470    
471     matches a portion of a subject string that is identical to itself. The power of
472     regular expressions comes from the ability to include alternatives and
473     repetitions in the pattern. These are encoded in the pattern by the use of
474     \fImeta-characters\fR, which do not stand for themselves but instead are
475     interpreted in some special way.
476    
477     There are two different sets of meta-characters: those that are recognized
478     anywhere in the pattern except within square brackets, and those that are
479     recognized in square brackets. Outside square brackets, the meta-characters are
480     as follows:
481    
482     \\ general escape character with several uses
483     ^ assert start of subject (or line, in multiline mode)
484     $ assert end of subject (or line, in multiline mode)
485     . match any character except newline (by default)
486     [ start character class definition
487     | start of alternative branch
488     ( start subpattern
489     ) end subpattern
490     ? extends the meaning of (
491     also 0 or 1 quantifier
492     also quantifier minimizer
493     * 0 or more quantifier
494     + 1 or more quantifier
495     { start min/max quantifier
496    
497     Part of a pattern that is in square brackets is called a "character class". In
498     a character class the only meta-characters are:
499    
500     \\ general escape character
501     ^ negate the class, but only if the first character
502     - indicates character range
503     ] terminates the character class
504    
505     The following sections describe the use of each of the meta-characters.
506    
507    
508     .SH BACKSLASH
509     The backslash character has several uses. Firstly, if it is followed by a
510     non-alphameric character, it takes away any special meaning that character may
511     have. This use of backslash as an escape character applies both inside and
512     outside character classes.
513    
514     For example, if you want to match a "*" character, you write "\\*" in the
515     pattern. This applies whether or not the following character would otherwise be
516     interpreted as a meta-character, so it is always safe to precede a
517     non-alphameric with "\\" to specify that it stands for itself. In particular,
518     if you want to match a backslash, you write "\\\\".
519    
520     If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
521 nigel 23 pattern (other than in a character class) and characters between a "#" outside
522     a character class and the next newline character are ignored. An escaping
523     backslash can be used to include a whitespace or "#" character as part of the
524     pattern.
525 nigel 3
526     A second use of backslash provides a way of encoding non-printing characters
527     in patterns in a visible manner. There is no restriction on the appearance of
528     non-printing characters, apart from the binary zero that terminates a pattern,
529     but when a pattern is being prepared by text editing, it is usually easier to
530     use one of the following escape sequences than the binary character it
531     represents:
532    
533     \\a alarm, that is, the BEL character (hex 07)
534     \\cx "control-x", where x is any character
535     \\e escape (hex 1B)
536     \\f formfeed (hex 0C)
537     \\n newline (hex 0A)
538     \\r carriage return (hex 0D)
539     \\t tab (hex 09)
540     \\xhh character with hex code hh
541 nigel 23 \\ddd character with octal code ddd, or backreference
542 nigel 3
543     The precise effect of "\\cx" is as follows: if "x" is a lower case letter, it
544     is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
545     Thus "\\cz" becomes hex 1A, but "\\c{" becomes hex 3B, while "\\c;" becomes hex
546     7B.
547    
548     After "\\x", up to two hexadecimal digits are read (letters can be in upper or
549     lower case).
550    
551     After "\\0" up to two further octal digits are read. In both cases, if there
552     are fewer than two digits, just those that are present are used. Thus the
553     sequence "\\0\\x\\07" specifies two binary zeros followed by a BEL character.
554 nigel 23 Make sure you supply two digits after the initial zero if the character that
555     follows is itself an octal digit.
556 nigel 3
557     The handling of a backslash followed by a digit other than 0 is complicated.
558     Outside a character class, PCRE reads it and any following digits as a decimal
559     number. If the number is less than 10, or if there have been at least that many
560     previous capturing left parentheses in the expression, the entire sequence is
561     taken as a \fIback reference\fR. A description of how this works is given
562     later, following the discussion of parenthesized subpatterns.
563    
564     Inside a character class, or if the decimal number is greater than 9 and there
565     have not been that many capturing subpatterns, PCRE re-reads up to three octal
566     digits following the backslash, and generates a single byte from the least
567     significant 8 bits of the value. Any subsequent digits stand for themselves.
568     For example:
569    
570     \\040 is another way of writing a space
571     \\40 is the same, provided there are fewer than 40
572     previous capturing subpatterns
573     \\7 is always a back reference
574     \\11 might be a back reference, or another way of
575     writing a tab
576     \\011 is always a tab
577     \\0113 is a tab followed by the character "3"
578     \\113 is the character with octal code 113 (since there
579     can be no more than 99 back references)
580     \\377 is a byte consisting entirely of 1 bits
581     \\81 is either a back reference, or a binary zero
582     followed by the two characters "8" and "1"
583    
584     Note that octal values of 100 or greater must not be introduced by a leading
585     zero, because no more than three octal digits are ever read.
586    
587     All the sequences that define a single byte value can be used both inside and
588     outside character classes. In addition, inside a character class, the sequence
589     "\\b" is interpreted as the backspace character (hex 08). Outside a character
590     class it has a different meaning (see below).
591    
592     The third use of backslash is for specifying generic character types:
593    
594     \\d any decimal digit
595     \\D any character that is not a decimal digit
596     \\s any whitespace character
597     \\S any character that is not a whitespace character
598     \\w any "word" character
599     \\W any "non-word" character
600    
601     Each pair of escape sequences partitions the complete set of characters into
602     two disjoint sets. Any given character matches one, and only one, of each pair.
603    
604     A "word" character is any letter or digit or the underscore character, that is,
605 nigel 25 any character which can be part of a Perl "word". The definition of letters and
606     digits is controlled by PCRE's character tables, and may vary if locale-
607     specific matching is taking place (see "Locale support" above). For example, in
608     the "fr" (French) locale, some character codes greater than 128 are used for
609     accented letters, and these are matched by \\w.
610 nigel 3
611 nigel 25 These character type sequences can appear both inside and outside character
612     classes. They each match one character of the appropriate type. If the current
613     matching point is at the end of the subject string, all of them fail, since
614     there is no character to match.
615    
616 nigel 23 The fourth use of backslash is for certain simple assertions. An assertion
617     specifies a condition that has to be met at a particular point in a match,
618     without consuming any characters from the subject string. The use of
619     subpatterns for more complicated assertions is described below. The backslashed
620     assertions are
621 nigel 3
622     \\b word boundary
623     \\B not a word boundary
624     \\A start of subject (independent of multiline mode)
625 nigel 23 \\Z end of subject or newline at end (independent of multiline mode)
626     \\z end of subject (independent of multiline mode)
627 nigel 3
628 nigel 23 These assertions may not appear in character classes (but note that "\\b" has a
629 nigel 3 different meaning, namely the backspace character, inside a character class).
630    
631     A word boundary is a position in the subject string where the current character
632 nigel 23 and the previous character do not both match \\w or \\W (i.e. one matches
633     \\w and the other matches \\W), or the start or end of the string if the
634     first or last character matches \\w, respectively.
635 nigel 3
636 nigel 23 The \\A, \\Z, and \\z assertions differ from the traditional circumflex and
637     dollar (described below) in that they only ever match at the very start and end
638     of the subject string, whatever options are set. They are not affected by the
639     PCRE_NOTBOL or PCRE_NOTEOL options. The difference between \\Z and \\z is that
640     \\Z matches before a newline that is the last character of the string as well
641     as at the end of the string, whereas \\z matches only at the end.
642 nigel 3
643    
644     .SH CIRCUMFLEX AND DOLLAR
645 nigel 23 Outside a character class, in the default matching mode, the circumflex
646     character is an assertion which is true only if the current matching point is
647     at the start of the subject string. Inside a character class, circumflex has an
648 nigel 3 entirely different meaning (see below).
649    
650     Circumflex need not be the first character of the pattern if a number of
651     alternatives are involved, but it should be the first thing in each alternative
652     in which it appears if the pattern is ever to match that branch. If all
653     possible alternatives start with a circumflex, that is, if the pattern is
654     constrained to match only at the start of the subject, it is said to be an
655     "anchored" pattern. (There are also other constructs that can cause a pattern
656     to be anchored.)
657    
658     A dollar character is an assertion which is true only if the current matching
659     point is at the end of the subject string, or immediately before a newline
660     character that is the last character in the string (by default). Dollar need
661     not be the last character of the pattern if a number of alternatives are
662     involved, but it should be the last item in any branch in which it appears.
663     Dollar has no special meaning in a character class.
664    
665     The meaning of dollar can be changed so that it matches only at the very end of
666     the string, by setting the PCRE_DOLLAR_ENDONLY option at compile or matching
667 nigel 23 time. This does not affect the \\Z assertion.
668 nigel 3
669     The meanings of the circumflex and dollar characters are changed if the
670 nigel 23 PCRE_MULTILINE option is set. When this is the case, they match immediately
671     after and immediately before an internal "\\n" character, respectively, in
672     addition to matching at the start and end of the subject string. For example,
673     the pattern /^abc$/ matches the subject string "def\\nabc" in multiline mode,
674     but not otherwise. Consequently, patterns that are anchored in single line mode
675     because all branches start with "^" are not anchored in multiline mode. The
676     PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
677 nigel 3
678 nigel 23 Note that the sequences \\A, \\Z, and \\z can be used to match the start and
679     end of the subject in both modes, and if all branches of a pattern start with
680     \\A is it always anchored, whether PCRE_MULTILINE is set or not.
681 nigel 3
682    
683     .SH FULL STOP (PERIOD, DOT)
684     Outside a character class, a dot in the pattern matches any one character in
685     the subject, including a non-printing character, but not (by default) newline.
686     If the PCRE_DOTALL option is set, then dots match newlines as well. The
687     handling of dot is entirely independent of the handling of circumflex and
688     dollar, the only relationship being that they both involve newline characters.
689     Dot has no special meaning in a character class.
690    
691    
692     .SH SQUARE BRACKETS
693     An opening square bracket introduces a character class, terminated by a closing
694     square bracket. A closing square bracket on its own is not special. If a
695     closing square bracket is required as a member of the class, it should be the
696     first data character in the class (after an initial circumflex, if present) or
697 nigel 23 escaped with a backslash.
698 nigel 3
699     A character class matches a single character in the subject; the character must
700     be in the set of characters defined by the class, unless the first character in
701     the class is a circumflex, in which case the subject character must not be in
702     the set defined by the class. If a circumflex is actually required as a member
703 nigel 23 of the class, ensure it is not the first character, or escape it with a
704     backslash.
705 nigel 3
706     For example, the character class [aeiou] matches any lower case vowel, while
707     [^aeiou] matches any character that is not a lower case vowel. Note that a
708     circumflex is just a convenient notation for specifying the characters which
709     are in the class by enumerating those that are not. It is not an assertion: it
710     still consumes a character from the subject string, and fails if the current
711     pointer is at the end of the string.
712    
713 nigel 25 When caseless matching is set, any letters in a class represent both their
714     upper case and lower case versions, so for example, a caseless [aeiou] matches
715     "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
716     caseful version would.
717 nigel 23
718 nigel 3 The newline character is never treated in any special way in character classes,
719     whatever the setting of the PCRE_DOTALL or PCRE_MULTILINE options is. A class
720     such as [^a] will always match a newline.
721    
722     The minus (hyphen) character can be used to specify a range of characters in a
723     character class. For example, [d-m] matches any letter between d and m,
724     inclusive. If a minus character is required in a class, it must be escaped with
725 nigel 23 a backslash or appear in a position where it cannot be interpreted as
726     indicating a range, typically as the first or last character in the class. It
727     is not possible to have the character "]" as the end character of a range,
728     since a sequence such as [w-] is interpreted as a class of two characters. The
729     octal or hexadecimal representation of "]" can, however, be used to end a
730     range.
731 nigel 3
732     Ranges operate in ASCII collating sequence. They can also be used for
733 nigel 25 characters specified numerically, for example [\\000-\\037]. If a range that
734     includes letters is used when caseless matching is set, it matches the letters
735     in either case. For example, [W-c] is equivalent to [][\\^_`wxyzabc], matched
736     caselessly, and if character tables for the "fr" locale are in use,
737     [\\xc8-\\xcb] matches accented E characters in both cases.
738 nigel 3
739     The character types \\d, \\D, \\s, \\S, \\w, and \\W may also appear in a
740     character class, and add the characters that they match to the class. For
741 nigel 23 example, [\\dABCDEF] matches any hexadecimal digit. A circumflex can
742     conveniently be used with the upper case character types to specify a more
743     restricted set of characters than the matching lower case type. For example,
744     the class [^\\W_] matches any letter or digit, but not underscore.
745 nigel 3
746     All non-alphameric characters other than \\, -, ^ (at the start) and the
747     terminating ] are non-special in character classes, but it does no harm if they
748     are escaped.
749    
750    
751     .SH VERTICAL BAR
752 nigel 23 Vertical bar characters are used to separate alternative patterns. For example,
753     the pattern
754 nigel 3
755     gilbert|sullivan
756    
757 nigel 23 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
758 nigel 3 and an empty alternative is permitted (matching the empty string).
759 nigel 23 The matching process tries each alternative in turn, from left to right,
760     and the first one that succeeds is used. If the alternatives are within a
761     subpattern (defined below), "succeeds" means matching the rest of the main
762     pattern as well as the alternative in the subpattern.
763 nigel 3
764    
765 nigel 23 .SH INTERNAL OPTION SETTING
766     The settings of PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and PCRE_EXTENDED
767     can be changed from within the pattern by a sequence of Perl option letters
768     enclosed between "(?" and ")". The option letters are
769    
770     i for PCRE_CASELESS
771     m for PCRE_MULTILINE
772     s for PCRE_DOTALL
773     x for PCRE_EXTENDED
774    
775     For example, (?im) sets caseless, multiline matching. It is also possible to
776     unset these options by preceding the letter with a hyphen, and a combined
777     setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
778     PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
779     permitted. If a letter appears both before and after the hyphen, the option is
780     unset.
781    
782     The scope of these option changes depends on where in the pattern the setting
783     occurs. For settings that are outside any subpattern (defined below), the
784     effect is the same as if the options were set or unset at the start of
785     matching. The following patterns all behave in exactly the same way:
786    
787     (?i)abc
788     a(?i)bc
789     ab(?i)c
790     abc(?i)
791    
792     which in turn is the same as compiling the pattern abc with PCRE_CASELESS set.
793     In other words, such "top level" settings apply to the whole pattern (unless
794     there are other changes inside subpatterns). If there is more than one setting
795     of the same option at top level, the rightmost setting is used.
796    
797     If an option change occurs inside a subpattern, the effect is different. This
798     is a change of behaviour in Perl 5.005. An option change inside a subpattern
799     affects only that part of the subpattern that follows it, so
800    
801     (a(?i)b)c
802    
803     matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
804     By this means, options can be made to have different settings in different
805     parts of the pattern. Any changes made in one alternative do carry on
806     into subsequent branches within the same subpattern. For example,
807    
808     (a(?i)b|c)
809    
810     matches "ab", "aB", "c", and "C", even though when matching "C" the first
811     branch is abandoned before the option setting. This is because the effects of
812     option settings happen at compile time. There would be some very weird
813     behaviour otherwise.
814    
815     The PCRE-specific options PCRE_UNGREEDY and PCRE_EXTRA can be changed in the
816     same way as the Perl-compatible options by using the characters U and X
817     respectively. The (?X) flag setting is special in that it must always occur
818     earlier in the pattern than any of the additional features it turns on, even
819     when it is at top level. It is best put at the start.
820    
821    
822 nigel 3 .SH SUBPATTERNS
823     Subpatterns are delimited by parentheses (round brackets), which can be nested.
824     Marking part of a pattern as a subpattern does two things:
825    
826     1. It localizes a set of alternatives. For example, the pattern
827    
828     cat(aract|erpillar|)
829    
830     matches one of the words "cat", "cataract", or "caterpillar". Without the
831     parentheses, it would match "cataract", "erpillar" or the empty string.
832    
833     2. It sets up the subpattern as a capturing subpattern (as defined above).
834     When the whole pattern matches, that portion of the subject string that matched
835     the subpattern is passed back to the caller via the \fIovector\fR argument of
836     \fBpcre_exec()\fR. Opening parentheses are counted from left to right (starting
837     from 1) to obtain the numbers of the capturing subpatterns.
838    
839     For example, if the string "the red king" is matched against the pattern
840    
841     the ((red|white) (king|queen))
842    
843     the captured substrings are "red king", "red", and "king", and are numbered 1,
844     2, and 3.
845    
846     The fact that plain parentheses fulfil two functions is not always helpful.
847     There are often times when a grouping subpattern is required without a
848     capturing requirement. If an opening parenthesis is followed by "?:", the
849     subpattern does not do any capturing, and is not counted when computing the
850     number of any subsequent capturing subpatterns. For example, if the string "the
851     white queen" is matched against the pattern
852    
853     the ((?:red|white) (king|queen))
854    
855     the captured substrings are "white queen" and "queen", and are numbered 1 and
856     2. The maximum number of captured substrings is 99, and the maximum number of
857     all subpatterns, both capturing and non-capturing, is 200.
858    
859 nigel 23 As a convenient shorthand, if any option settings are required at the start of
860     a non-capturing subpattern, the option letters may appear between the "?" and
861     the ":". Thus the two patterns
862 nigel 3
863 nigel 23 (?i:saturday|sunday)
864     (?:(?i)saturday|sunday)
865 nigel 3
866 nigel 23 match exactly the same set of strings. Because alternative branches are tried
867     from left to right, and options are not reset until the end of the subpattern
868     is reached, an option setting in one branch does affect subsequent branches, so
869     the above patterns match "SUNDAY" as well as "Saturday".
870 nigel 3
871    
872     .SH REPETITION
873     Repetition is specified by quantifiers, which can follow any of the following
874     items:
875    
876     a single character, possibly escaped
877     the . metacharacter
878     a character class
879 nigel 23 a back reference (see next section)
880     a parenthesized subpattern (unless it is an assertion - see below)
881 nigel 3
882     The general repetition quantifier specifies a minimum and maximum number of
883     permitted matches, by giving the two numbers in curly brackets (braces),
884     separated by a comma. The numbers must be less than 65536, and the first must
885     be less than or equal to the second. For example:
886    
887     z{2,4}
888    
889     matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
890     character. If the second number is omitted, but the comma is present, there is
891     no upper limit; if the second number and the comma are both omitted, the
892     quantifier specifies an exact number of required matches. Thus
893    
894     [aeiou]{3,}
895    
896     matches at least 3 successive vowels, but may match many more, while
897    
898     \\d{8}
899    
900     matches exactly 8 digits. An opening curly bracket that appears in a position
901     where a quantifier is not allowed, or one that does not match the syntax of a
902 nigel 23 quantifier, is taken as a literal character. For example, {,6} is not a
903 nigel 3 quantifier, but a literal string of four characters.
904    
905     The quantifier {0} is permitted, causing the expression to behave as if the
906     previous item and the quantifier were not present.
907    
908     For convenience (and historical compatibility) the three most common
909     quantifiers have single-character abbreviations:
910    
911     * is equivalent to {0,}
912     + is equivalent to {1,}
913     ? is equivalent to {0,1}
914    
915 nigel 23 It is possible to construct infinite loops by following a subpattern that can
916     match no characters with a quantifier that has no upper limit, for example:
917    
918     (a?)*
919    
920     Earlier versions of Perl and PCRE used to give an error at compile time for
921     such patterns. However, because there are cases where this can be useful, such
922     patterns are now accepted, but if any repetition of the subpattern does in fact
923     match no characters, the loop is forcibly broken.
924    
925 nigel 3 By default, the quantifiers are "greedy", that is, they match as much as
926     possible (up to the maximum number of permitted times), without causing the
927     rest of the pattern to fail. The classic example of where this gives problems
928     is in trying to match comments in C programs. These appear between the
929     sequences /* and */ and within the sequence, individual * and / characters may
930     appear. An attempt to match C comments by applying the pattern
931    
932     /\\*.*\\*/
933    
934     to the string
935    
936     /* first command */ not comment /* second comment */
937    
938     fails, because it matches the entire string due to the greediness of the .*
939     item.
940    
941     However, if a quantifier is followed by a question mark, then it ceases to be
942     greedy, and instead matches the minimum number of times possible, so the
943     pattern
944    
945     /\\*.*?\\*/
946    
947     does the right thing with the C comments. The meaning of the various
948     quantifiers is not otherwise changed, just the preferred number of matches.
949     Do not confuse this use of question mark with its use as a quantifier in its
950     own right. Because it has two uses, it can sometimes appear doubled, as in
951    
952 nigel 23 \\d??\\d
953 nigel 3
954     which matches one digit by preference, but can match two if that is the only
955     way the rest of the pattern matches.
956    
957 nigel 19 If the PCRE_UNGREEDY option is set (an option which is not available in Perl)
958     then the quantifiers are not greedy by default, but individual ones can be made
959 nigel 23 greedy by following them with a question mark. In other words, it inverts the
960 nigel 19 default behaviour.
961    
962 nigel 7 When a parenthesized subpattern is quantified with a minimum repeat count that
963 nigel 3 is greater than 1 or with a limited maximum, more store is required for the
964     compiled pattern, in proportion to the size of the minimum or maximum.
965    
966     If a pattern starts with .* then it is implicitly anchored, since whatever
967     follows will be tried against every character position in the subject string.
968     PCRE treats this as though it were preceded by \\A.
969    
970     When a capturing subpattern is repeated, the value captured is the substring
971 nigel 23 that matched the final iteration. For example, after
972 nigel 3
973 nigel 23 (tweedle[dume]{3}\\s*)+
974 nigel 3
975 nigel 23 has matched "tweedledum tweedledee" the value of the captured substring is
976     "tweedledee". However, if there are nested capturing subpatterns, the
977     corresponding captured values may have been set in previous iterations. For
978     example, after
979 nigel 3
980 nigel 23 /(a|(b))+/
981 nigel 3
982 nigel 23 matches "aba" the value of the second captured substring is "b".
983    
984    
985     .SH BACK REFERENCES
986     Outside a character class, a backslash followed by a digit greater than 0 (and
987     possibly further digits) is a back reference to a capturing subpattern earlier
988     (i.e. to its left) in the pattern, provided there have been that many previous
989     capturing left parentheses.
990    
991     However, if the decimal number following the backslash is less than 10, it is
992     always taken as a back reference, and causes an error only if there are not
993     that many capturing left parentheses in the entire pattern. In other words, the
994     parentheses that are referenced need not be to the left of the reference for
995     numbers less than 10. See the section entitled "Backslash" above for further
996     details of the handling of digits following a backslash.
997    
998     A back reference matches whatever actually matched the capturing subpattern in
999     the current subject string, rather than anything matching the subpattern
1000     itself. So the pattern
1001    
1002     (sens|respons)e and \\1ibility
1003    
1004     matches "sense and sensibility" and "response and responsibility", but not
1005     "sense and responsibility". If caseful matching is in force at the time of the
1006     back reference, then the case of letters is relevant. For example,
1007    
1008     ((?i)rah)\\s+\\1
1009    
1010     matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1011     capturing subpattern is matched caselessly.
1012    
1013     There may be more than one back reference to the same subpattern. If a
1014     subpattern has not actually been used in a particular match, then any back
1015     references to it always fail. For example, the pattern
1016    
1017     (a|(bc))\\2
1018    
1019     always fails if it starts to match "a" rather than "bc". Because there may be
1020     up to 99 back references, all digits following the backslash are taken
1021     as part of a potential back reference number. If the pattern continues with a
1022     digit character, then some delimiter must be used to terminate the back
1023     reference. If the PCRE_EXTENDED option is set, this can be whitespace.
1024     Otherwise an empty comment can be used.
1025    
1026     A back reference that occurs inside the parentheses to which it refers fails
1027     when the subpattern is first used, so, for example, (a\\1) never matches.
1028     However, such references can be useful inside repeated subpatterns. For
1029     example, the pattern
1030    
1031     (a|b\\1)+
1032    
1033     matches any number of "a"s and also "aba", "ababaa" etc. At each iteration of
1034     the subpattern, the back reference matches the character string corresponding
1035     to the previous iteration. In order for this to work, the pattern must be such
1036     that the first iteration does not need to match the back reference. This can be
1037     done using alternation, as in the example above, or by a quantifier with a
1038     minimum of zero.
1039    
1040    
1041 nigel 3 .SH ASSERTIONS
1042 nigel 23 An assertion is a test on the characters following or preceding the current
1043     matching point that does not actually consume any characters. The simple
1044     assertions coded as \\b, \\B, \\A, \\Z, \\z, ^ and $ are described above. More
1045     complicated assertions are coded as subpatterns. There are two kinds: those
1046     that look ahead of the current position in the subject string, and those that
1047     look behind it.
1048 nigel 3
1049 nigel 23 An assertion subpattern is matched in the normal way, except that it does not
1050     cause the current matching position to be changed. Lookahead assertions start
1051     with (?= for positive assertions and (?! for negative assertions. For example,
1052    
1053 nigel 3 \\w+(?=;)
1054    
1055     matches a word followed by a semicolon, but does not include the semicolon in
1056     the match, and
1057    
1058     foo(?!bar)
1059    
1060     matches any occurrence of "foo" that is not followed by "bar". Note that the
1061     apparently similar pattern
1062    
1063     (?!foo)bar
1064    
1065     does not find an occurrence of "bar" that is preceded by something other than
1066     "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1067 nigel 23 (?!foo) is always true when the next three characters are "bar". A
1068     lookbehind assertion is needed to achieve this effect.
1069 nigel 3
1070 nigel 23 Lookbehind assertions start with (?<= for positive assertions and (?<! for
1071     negative assertions. For example,
1072    
1073     (?<!foo)bar
1074    
1075     does find an occurrence of "bar" that is not preceded by "foo". The contents of
1076     a lookbehind assertion are restricted such that all the strings it matches must
1077     have a fixed length. However, if there are several alternatives, they do not
1078     all have to have the same fixed length. Thus
1079    
1080     (?<=bullock|donkey)
1081    
1082     is permitted, but
1083    
1084     (?<!dogs?|cats?)
1085    
1086     causes an error at compile time. Branches that match different length strings
1087     are permitted only at the top level of a lookbehind assertion. This is an
1088     extension compared with Perl 5.005, which requires all branches to match the
1089     same length of string. An assertion such as
1090    
1091     (?<=ab(c|de))
1092    
1093 nigel 27 is not permitted, because its single top-level branch can match two different
1094     lengths, but it is acceptable if rewritten to use two top-level branches:
1095 nigel 23
1096     (?<=abc|abde)
1097    
1098     The implementation of lookbehind assertions is, for each alternative, to
1099     temporarily move the current position back by the fixed width and then try to
1100     match. If there are insufficient characters before the current position, the
1101 nigel 27 match is deemed to fail. Lookbehinds in conjunction with once-only subpatterns
1102     can be particularly useful for matching at the ends of strings; an example is
1103     given at the end of the section on once-only subpatterns.
1104 nigel 23
1105 nigel 27 Several assertions (of any sort) may occur in succession. For example,
1106 nigel 23
1107 nigel 27 (?<=\\d{3})(?<!999)foo
1108    
1109     matches "foo" preceded by three digits that are not "999". Furthermore,
1110     assertions can be nested in any combination. For example,
1111    
1112 nigel 23 (?<=(?<!foo)bar)baz
1113    
1114     matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1115     preceded by "foo".
1116    
1117 nigel 3 Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1118     because it makes no sense to assert the same thing several times. If an
1119     assertion contains capturing subpatterns within it, these are always counted
1120     for the purposes of numbering the capturing subpatterns in the whole pattern.
1121     Substring capturing is carried out for positive assertions, but it does not
1122     make sense for negative assertions.
1123    
1124     Assertions count towards the maximum of 200 parenthesized subpatterns.
1125    
1126    
1127     .SH ONCE-ONLY SUBPATTERNS
1128     With both maximizing and minimizing repetition, failure of what follows
1129     normally causes the repeated item to be re-evaluated to see if a different
1130     number of repeats allows the rest of the pattern to match. Sometimes it is
1131     useful to prevent this, either to change the nature of the match, or to cause
1132 nigel 23 it fail earlier than it otherwise might, when the author of the pattern knows
1133 nigel 3 there is no point in carrying on.
1134    
1135     Consider, for example, the pattern \\d+foo when applied to the subject line
1136    
1137 nigel 23 123456bar
1138 nigel 3
1139     After matching all 6 digits and then failing to match "foo", the normal
1140     action of the matcher is to try again with only 5 digits matching the \\d+
1141     item, and then with 4, and so on, before ultimately failing. Once-only
1142     subpatterns provide the means for specifying that once a portion of the pattern
1143     has matched, it is not to be re-evaluated in this way, so the matcher would
1144     give up immediately on failing to match "foo" the first time. The notation is
1145     another kind of special parenthesis, starting with (?> as in this example:
1146    
1147 nigel 23 (?>\\d+)bar
1148 nigel 3
1149     This kind of parenthesis "locks up" the part of the pattern it contains once
1150     it has matched, and a failure further into the pattern is prevented from
1151     backtracking into it. Backtracking past it to previous items, however, works as
1152     normal.
1153    
1154 nigel 23 An alternative description is that a subpattern of this type matches the string
1155     of characters that an identical standalone pattern would match, if anchored at
1156     the current point in the subject string.
1157 nigel 3
1158 nigel 23 Once-only subpatterns are not capturing subpatterns. Simple cases such as the
1159     above example can be though of as a maximizing repeat that must swallow
1160     everything it can. So, while both \\d+ and \\d+? are prepared to adjust the
1161     number of digits they match in order to make the rest of the pattern match,
1162     (?>\\d+) can only match an entire sequence of digits.
1163    
1164 nigel 3 This construction can of course contain arbitrarily complicated subpatterns,
1165 nigel 23 and it can be nested.
1166 nigel 3
1167 nigel 27 Once-only subpatterns can be used in conjunction with lookbehind assertions to
1168     specify efficient matching at the end of the subject string. Consider a simple
1169     pattern such as
1170 nigel 3
1171 nigel 27 abcd$
1172    
1173     when applied to a long string which does not match it. Because matching
1174     proceeds from left to right, PCRE will look for each "a" in the subject and
1175     then see if what follows matches the rest of the pattern. If the pattern is
1176     specified as
1177    
1178     .*abcd$
1179    
1180     then the initial .* matches the entire string at first, but when this fails, it
1181     backtracks to match all but the last character, then all but the last two
1182     characters, and so on. Once again the search for "a" covers the entire string,
1183     from right to left, so we are no better off. However, if the pattern is written
1184     as
1185    
1186     (?>.*)(?<=abcd)
1187    
1188     then there can be no backtracking for the .* item; it can match only the entire
1189     string. The subsequent lookbehind assertion does a single test on the last four
1190     characters. If it fails, the match fails immediately. For long strings, this
1191     approach makes a significant difference to the processing time.
1192    
1193    
1194 nigel 23 .SH CONDITIONAL SUBPATTERNS
1195     It is possible to cause the matching process to obey a subpattern
1196     conditionally or to choose between two alternative subpatterns, depending on
1197     the result of an assertion, or whether a previous capturing subpattern matched
1198     or not. The two possible forms of conditional subpattern are
1199 nigel 3
1200 nigel 23 (?(condition)yes-pattern)
1201     (?(condition)yes-pattern|no-pattern)
1202 nigel 3
1203 nigel 23 If the condition is satisfied, the yes-pattern is used; otherwise the
1204     no-pattern (if present) is used. If there are more than two alternatives in the
1205     subpattern, a compile-time error occurs.
1206 nigel 3
1207 nigel 23 There are two kinds of condition. If the text between the parentheses consists
1208     of a sequence of digits, then the condition is satisfied if the capturing
1209     subpattern of that number has previously matched. Consider the following
1210     pattern, which contains non-significant white space to make it more readable
1211     (assume the PCRE_EXTENDED option) and to divide it into three parts for ease
1212     of discussion:
1213 nigel 3
1214 nigel 23 ( \\( )? [^()]+ (?(1) \\) )
1215 nigel 3
1216 nigel 23 The first part matches an optional opening parenthesis, and if that
1217     character is present, sets it as the first captured substring. The second part
1218     matches one or more characters that are not parentheses. The third part is a
1219     conditional subpattern that tests whether the first set of parentheses matched
1220     or not. If they did, that is, if subject started with an opening parenthesis,
1221     the condition is true, and so the yes-pattern is executed and a closing
1222     parenthesis is required. Otherwise, since no-pattern is not present, the
1223     subpattern matches nothing. In other words, this pattern matches a sequence of
1224     non-parentheses, optionally enclosed in parentheses.
1225 nigel 3
1226 nigel 23 If the condition is not a sequence of digits, it must be an assertion. This may
1227     be a positive or negative lookahead or lookbehind assertion. Consider this
1228     pattern, again containing non-significant white space, and with the two
1229     alternatives on the second line:
1230 nigel 3
1231 nigel 23 (?(?=[^a-z]*[a-z])
1232     \\d{2}[a-z]{3}-\\d{2} | \\d{2}-\\d{2}-\\d{2} )
1233 nigel 19
1234 nigel 23 The condition is a positive lookahead assertion that matches an optional
1235     sequence of non-letters followed by a letter. In other words, it tests for the
1236     presence of at least one letter in the subject. If a letter is found, the
1237     subject is matched against the first alternative; otherwise it is matched
1238     against the second. This pattern matches strings in one of the two forms
1239     dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
1240 nigel 19
1241 nigel 3
1242 nigel 23 .SH COMMENTS
1243     The sequence (?# marks the start of a comment which continues up to the next
1244     closing parenthesis. Nested parentheses are not permitted. The characters
1245     that make up a comment play no part in the pattern matching at all.
1246 nigel 3
1247 nigel 23 If the PCRE_EXTENDED option is set, an unescaped # character outside a
1248     character class introduces a comment that continues up to the next newline
1249     character in the pattern.
1250    
1251    
1252 nigel 3 .SH PERFORMANCE
1253     Certain items that may appear in patterns are more efficient than others. It is
1254     more efficient to use a character class like [aeiou] than a set of alternatives
1255     such as (a|e|i|o|u). In general, the simplest construction that provides the
1256     required behaviour is usually the most efficient. Jeffrey Friedl's book
1257     contains a lot of discussion about optimizing regular expressions for efficient
1258     performance.
1259    
1260    
1261     .SH AUTHOR
1262     Philip Hazel <ph10@cam.ac.uk>
1263     .br
1264     University Computing Service,
1265     .br
1266     New Museums Site,
1267     .br
1268     Cambridge CB2 3QG, England.
1269     .br
1270     Phone: +44 1223 334714
1271    
1272 nigel 27 Copyright (c) 1997-1999 University of Cambridge.

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12