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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 51 - (show annotations) (download)
Sat Feb 24 21:39:37 2007 UTC (7 years, 2 months ago) by nigel
File size: 83884 byte(s)
Load pcre-3.4 into code/trunk.

1 .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 .B const char **\fIerrptr\fR, int *\fIerroffset\fR,
12 .ti +5n
13 .B const unsigned char *\fItableptr\fR);
14 .PP
15 .br
16 .B pcre_extra *pcre_study(const pcre *\fIcode\fR, int \fIoptions\fR,
17 .ti +5n
18 .B const char **\fIerrptr\fR);
19 .PP
20 .br
21 .B int pcre_exec(const pcre *\fIcode\fR, "const pcre_extra *\fIextra\fR,"
22 .ti +5n
23 .B "const char *\fIsubject\fR," int \fIlength\fR, int \fIstartoffset\fR,
24 .ti +5n
25 .B int \fIoptions\fR, int *\fIovector\fR, int \fIovecsize\fR);
26 .PP
27 .br
28 .B int pcre_copy_substring(const char *\fIsubject\fR, int *\fIovector\fR,
29 .ti +5n
30 .B int \fIstringcount\fR, int \fIstringnumber\fR, char *\fIbuffer\fR,
31 .ti +5n
32 .B int \fIbuffersize\fR);
33 .PP
34 .br
35 .B int pcre_get_substring(const char *\fIsubject\fR, int *\fIovector\fR,
36 .ti +5n
37 .B int \fIstringcount\fR, int \fIstringnumber\fR,
38 .ti +5n
39 .B const char **\fIstringptr\fR);
40 .PP
41 .br
42 .B int pcre_get_substring_list(const char *\fIsubject\fR,
43 .ti +5n
44 .B int *\fIovector\fR, int \fIstringcount\fR, "const char ***\fIlistptr\fR);"
45 .PP
46 .br
47 .B void pcre_free_substring(const char *\fIstringptr\fR);
48 .PP
49 .br
50 .B void pcre_free_substring_list(const char **\fIstringptr\fR);
51 .PP
52 .br
53 .B const unsigned char *pcre_maketables(void);
54 .PP
55 .br
56 .B int pcre_fullinfo(const pcre *\fIcode\fR, "const pcre_extra *\fIextra\fR,"
57 .ti +5n
58 .B int \fIwhat\fR, void *\fIwhere\fR);
59 .PP
60 .br
61 .B int pcre_info(const pcre *\fIcode\fR, int *\fIoptptr\fR, int
62 .B *\fIfirstcharptr\fR);
63 .PP
64 .br
65 .B char *pcre_version(void);
66 .PP
67 .br
68 .B void *(*pcre_malloc)(size_t);
69 .PP
70 .br
71 .B void (*pcre_free)(void *);
72
73
74
75 .SH DESCRIPTION
76 The PCRE library is a set of functions that implement regular expression
77 pattern matching using the same syntax and semantics as Perl 5, with just a few
78 differences (see below). The current implementation corresponds to Perl 5.005,
79 with some additional features from later versions. This includes some
80 experimental, incomplete support for UTF-8 encoded strings. Details of exactly
81 what is and what is not supported are given below.
82
83 PCRE has its own native API, which is described in this document. There is also
84 a set of wrapper functions that correspond to the POSIX regular expression API.
85 These are described in the \fBpcreposix\fR documentation.
86
87 The native API function prototypes are defined in the header file \fBpcre.h\fR,
88 and on Unix systems the library itself is called \fBlibpcre.a\fR, so can be
89 accessed by adding \fB-lpcre\fR to the command for linking an application which
90 calls it. The header file defines the macros PCRE_MAJOR and PCRE_MINOR to
91 contain the major and minor release numbers for the library. Applications can
92 use these to include support for different releases.
93
94 The functions \fBpcre_compile()\fR, \fBpcre_study()\fR, and \fBpcre_exec()\fR
95 are used for compiling and matching regular expressions.
96
97 The functions \fBpcre_copy_substring()\fR, \fBpcre_get_substring()\fR, and
98 \fBpcre_get_substring_list()\fR are convenience functions for extracting
99 captured substrings from a matched subject string; \fBpcre_free_substring()\fR
100 and \fBpcre_free_substring_list()\fR are also provided, to free the memory used
101 for extracted strings.
102
103 The function \fBpcre_maketables()\fR is used (optionally) to build a set of
104 character tables in the current locale for passing to \fBpcre_compile()\fR.
105
106 The function \fBpcre_fullinfo()\fR is used to find out information about a
107 compiled pattern; \fBpcre_info()\fR is an obsolete version which returns only
108 some of the available information, but is retained for backwards compatibility.
109 The function \fBpcre_version()\fR returns a pointer to a string containing the
110 version of PCRE and its date of release.
111
112 The global variables \fBpcre_malloc\fR and \fBpcre_free\fR initially contain
113 the entry points of the standard \fBmalloc()\fR and \fBfree()\fR functions
114 respectively. PCRE calls the memory management functions via these variables,
115 so a calling program can replace them if it wishes to intercept the calls. This
116 should be done before calling any PCRE functions.
117
118
119 .SH MULTI-THREADING
120 The PCRE functions can be used in multi-threading applications, with the
121 proviso that the memory management functions pointed to by \fBpcre_malloc\fR
122 and \fBpcre_free\fR are shared by all threads.
123
124 The compiled form of a regular expression is not altered during matching, so
125 the same compiled pattern can safely be used by several threads at once.
126
127
128 .SH COMPILING A PATTERN
129 The function \fBpcre_compile()\fR is called to compile a pattern into an
130 internal form. The pattern is a C string terminated by a binary zero, and
131 is passed in the argument \fIpattern\fR. A pointer to a single block of memory
132 that is obtained via \fBpcre_malloc\fR is returned. This contains the
133 compiled code and related data. The \fBpcre\fR type is defined for this for
134 convenience, but in fact \fBpcre\fR is just a typedef for \fBvoid\fR, since the
135 contents of the block are not externally defined. It is up to the caller to
136 free the memory when it is no longer required.
137 .PP
138 The size of a compiled pattern is roughly proportional to the length of the
139 pattern string, except that each character class (other than those containing
140 just a single character, negated or not) requires 33 bytes, and repeat
141 quantifiers with a minimum greater than one or a bounded maximum cause the
142 relevant portions of the compiled pattern to be replicated.
143 .PP
144 The \fIoptions\fR argument contains independent bits that affect the
145 compilation. It should be zero if no options are required. Some of the options,
146 in particular, those that are compatible with Perl, can also be set and unset
147 from within the pattern (see the detailed description of regular expressions
148 below). For these options, the contents of the \fIoptions\fR argument specifies
149 their initial settings at the start of compilation and execution. The
150 PCRE_ANCHORED option can be set at the time of matching as well as at compile
151 time.
152 .PP
153 If \fIerrptr\fR is NULL, \fBpcre_compile()\fR returns NULL immediately.
154 Otherwise, if compilation of a pattern fails, \fBpcre_compile()\fR returns
155 NULL, and sets the variable pointed to by \fIerrptr\fR to point to a textual
156 error message. The offset from the start of the pattern to the character where
157 the error was discovered is placed in the variable pointed to by
158 \fIerroffset\fR, which must not be NULL. If it is, an immediate error is given.
159 .PP
160 If the final argument, \fItableptr\fR, is NULL, PCRE uses a default set of
161 character tables which are built when it is compiled, using the default C
162 locale. Otherwise, \fItableptr\fR must be the result of a call to
163 \fBpcre_maketables()\fR. See the section on locale support below.
164 .PP
165 The following option bits are defined in the header file:
166
167 PCRE_ANCHORED
168
169 If this bit is set, the pattern is forced to be "anchored", that is, it is
170 constrained to match only at the start of the string which is being searched
171 (the "subject string"). This effect can also be achieved by appropriate
172 constructs in the pattern itself, which is the only way to do it in Perl.
173
174 PCRE_CASELESS
175
176 If this bit is set, letters in the pattern match both upper and lower case
177 letters. It is equivalent to Perl's /i option.
178
179 PCRE_DOLLAR_ENDONLY
180
181 If this bit is set, a dollar metacharacter in the pattern matches only at the
182 end of the subject string. Without this option, a dollar also matches
183 immediately before the final character if it is a newline (but not before any
184 other newlines). The PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is
185 set. There is no equivalent to this option in Perl.
186
187 PCRE_DOTALL
188
189 If this bit is set, a dot metacharater in the pattern matches all characters,
190 including newlines. Without it, newlines are excluded. This option is
191 equivalent to Perl's /s option. A negative class such as [^a] always matches a
192 newline character, independent of the setting of this option.
193
194 PCRE_EXTENDED
195
196 If this bit is set, whitespace data characters in the pattern are totally
197 ignored except when escaped or inside a character class, and characters between
198 an unescaped # outside a character class and the next newline character,
199 inclusive, are also ignored. This is equivalent to Perl's /x option, and makes
200 it possible to include comments inside complicated patterns. Note, however,
201 that this applies only to data characters. Whitespace characters may never
202 appear within special character sequences in a pattern, for example within the
203 sequence (?( which introduces a conditional subpattern.
204
205 PCRE_EXTRA
206
207 This option was invented in order to turn on additional functionality of PCRE
208 that is incompatible with Perl, but it is currently of very little use. When
209 set, any backslash in a pattern that is followed by a letter that has no
210 special meaning causes an error, thus reserving these combinations for future
211 expansion. By default, as in Perl, a backslash followed by a letter with no
212 special meaning is treated as a literal. There are at present no other features
213 controlled by this option. It can also be set by a (?X) option setting within a
214 pattern.
215
216 PCRE_MULTILINE
217
218 By default, PCRE treats the subject string as consisting of a single "line" of
219 characters (even if it actually contains several newlines). The "start of line"
220 metacharacter (^) matches only at the start of the string, while the "end of
221 line" metacharacter ($) matches only at the end of the string, or before a
222 terminating newline (unless PCRE_DOLLAR_ENDONLY is set). This is the same as
223 Perl.
224
225 When PCRE_MULTILINE it is set, the "start of line" and "end of line" constructs
226 match immediately following or immediately before any newline in the subject
227 string, respectively, as well as at the very start and end. This is equivalent
228 to Perl's /m option. If there are no "\\n" characters in a subject string, or
229 no occurrences of ^ or $ in a pattern, setting PCRE_MULTILINE has no
230 effect.
231
232 PCRE_UNGREEDY
233
234 This option inverts the "greediness" of the quantifiers so that they are not
235 greedy by default, but become greedy if followed by "?". It is not compatible
236 with Perl. It can also be set by a (?U) option setting within the pattern.
237
238 PCRE_UTF8
239
240 This option causes PCRE to regard both the pattern and the subject as strings
241 of UTF-8 characters instead of just byte strings. However, it is available only
242 if PCRE has been built to include UTF-8 support. If not, the use of this option
243 provokes an error. Support for UTF-8 is new, experimental, and incomplete.
244 Details of exactly what it entails are given below.
245
246
247 .SH STUDYING A PATTERN
248 When a pattern is going to be used several times, it is worth spending more
249 time analyzing it in order to speed up the time taken for matching. The
250 function \fBpcre_study()\fR takes a pointer to a compiled pattern as its first
251 argument, and returns a pointer to a \fBpcre_extra\fR block (another \fBvoid\fR
252 typedef) containing additional information about the pattern; this can be
253 passed to \fBpcre_exec()\fR. If no additional information is available, NULL
254 is returned.
255
256 The second argument contains option bits. At present, no options are defined
257 for \fBpcre_study()\fR, and this argument should always be zero.
258
259 The third argument for \fBpcre_study()\fR is a pointer to an error message. If
260 studying succeeds (even if no data is returned), the variable it points to is
261 set to NULL. Otherwise it points to a textual error message.
262
263 At present, studying a pattern is useful only for non-anchored patterns that do
264 not have a single fixed starting character. A bitmap of possible starting
265 characters is created.
266
267
268 .SH LOCALE SUPPORT
269 PCRE handles caseless matching, and determines whether characters are letters,
270 digits, or whatever, by reference to a set of tables. The library contains a
271 default set of tables which is created in the default C locale when PCRE is
272 compiled. This is used when the final argument of \fBpcre_compile()\fR is NULL,
273 and is sufficient for many applications.
274
275 An alternative set of tables can, however, be supplied. Such tables are built
276 by calling the \fBpcre_maketables()\fR function, which has no arguments, in the
277 relevant locale. The result can then be passed to \fBpcre_compile()\fR as often
278 as necessary. For example, to build and use tables that are appropriate for the
279 French locale (where accented characters with codes greater than 128 are
280 treated as letters), the following code could be used:
281
282 setlocale(LC_CTYPE, "fr");
283 tables = pcre_maketables();
284 re = pcre_compile(..., tables);
285
286 The tables are built in memory that is obtained via \fBpcre_malloc\fR. The
287 pointer that is passed to \fBpcre_compile\fR is saved with the compiled
288 pattern, and the same tables are used via this pointer by \fBpcre_study()\fR
289 and \fBpcre_exec()\fR. Thus for any single pattern, compilation, studying and
290 matching all happen in the same locale, but different patterns can be compiled
291 in different locales. It is the caller's responsibility to ensure that the
292 memory containing the tables remains available for as long as it is needed.
293
294
295 .SH INFORMATION ABOUT A PATTERN
296 The \fBpcre_fullinfo()\fR function returns information about a compiled
297 pattern. It replaces the obsolete \fBpcre_info()\fR function, which is
298 nevertheless retained for backwards compability (and is documented below).
299
300 The first argument for \fBpcre_fullinfo()\fR is a pointer to the compiled
301 pattern. The second argument is the result of \fBpcre_study()\fR, or NULL if
302 the pattern was not studied. The third argument specifies which piece of
303 information is required, while the fourth argument is a pointer to a variable
304 to receive the data. The yield of the function is zero for success, or one of
305 the following negative numbers:
306
307 PCRE_ERROR_NULL the argument \fIcode\fR was NULL
308 the argument \fIwhere\fR was NULL
309 PCRE_ERROR_BADMAGIC the "magic number" was not found
310 PCRE_ERROR_BADOPTION the value of \fIwhat\fR was invalid
311
312 The possible values for the third argument are defined in \fBpcre.h\fR, and are
313 as follows:
314
315 PCRE_INFO_OPTIONS
316
317 Return a copy of the options with which the pattern was compiled. The fourth
318 argument should point to au \fBunsigned long int\fR variable. These option bits
319 are those specified in the call to \fBpcre_compile()\fR, modified by any
320 top-level option settings within the pattern itself, and with the PCRE_ANCHORED
321 bit forcibly set if the form of the pattern implies that it can match only at
322 the start of a subject string.
323
324 PCRE_INFO_SIZE
325
326 Return the size of the compiled pattern, that is, the value that was passed as
327 the argument to \fBpcre_malloc()\fR when PCRE was getting memory in which to
328 place the compiled data. The fourth argument should point to a \fBsize_t\fR
329 variable.
330
331 PCRE_INFO_CAPTURECOUNT
332
333 Return the number of capturing subpatterns in the pattern. The fourth argument
334 should point to an \fbint\fR variable.
335
336 PCRE_INFO_BACKREFMAX
337
338 Return the number of the highest back reference in the pattern. The fourth
339 argument should point to an \fBint\fR variable. Zero is returned if there are
340 no back references.
341
342 PCRE_INFO_FIRSTCHAR
343
344 Return information about the first character of any matched string, for a
345 non-anchored pattern. If there is a fixed first character, e.g. from a pattern
346 such as (cat|cow|coyote), it is returned in the integer pointed to by
347 \fIwhere\fR. Otherwise, if either
348
349 (a) the pattern was compiled with the PCRE_MULTILINE option, and every branch
350 starts with "^", or
351
352 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not set
353 (if it were set, the pattern would be anchored),
354
355 -1 is returned, indicating that the pattern matches only at the start of a
356 subject string or after any "\\n" within the string. Otherwise -2 is returned.
357 For anchored patterns, -2 is returned.
358
359 PCRE_INFO_FIRSTTABLE
360
361 If the pattern was studied, and this resulted in the construction of a 256-bit
362 table indicating a fixed set of characters for the first character in any
363 matching string, a pointer to the table is returned. Otherwise NULL is
364 returned. The fourth argument should point to an \fBunsigned char *\fR
365 variable.
366
367 PCRE_INFO_LASTLITERAL
368
369 For a non-anchored pattern, return the value of the rightmost literal character
370 which must exist in any matched string, other than at its start. The fourth
371 argument should point to an \fBint\fR variable. If there is no such character,
372 or if the pattern is anchored, -1 is returned. For example, for the pattern
373 /a\\d+z\\d+/ the returned value is 'z'.
374
375 The \fBpcre_info()\fR function is now obsolete because its interface is too
376 restrictive to return all the available data about a compiled pattern. New
377 programs should use \fBpcre_fullinfo()\fR instead. The yield of
378 \fBpcre_info()\fR is the number of capturing subpatterns, or one of the
379 following negative numbers:
380
381 PCRE_ERROR_NULL the argument \fIcode\fR was NULL
382 PCRE_ERROR_BADMAGIC the "magic number" was not found
383
384 If the \fIoptptr\fR argument is not NULL, a copy of the options with which the
385 pattern was compiled is placed in the integer it points to (see
386 PCRE_INFO_OPTIONS above).
387
388 If the pattern is not anchored and the \fIfirstcharptr\fR argument is not NULL,
389 it is used to pass back information about the first character of any matched
390 string (see PCRE_INFO_FIRSTCHAR above).
391
392
393 .SH MATCHING A PATTERN
394 The function \fBpcre_exec()\fR is called to match a subject string against a
395 pre-compiled pattern, which is passed in the \fIcode\fR argument. If the
396 pattern has been studied, the result of the study should be passed in the
397 \fIextra\fR argument. Otherwise this must be NULL.
398
399 The PCRE_ANCHORED option can be passed in the \fIoptions\fR argument, whose
400 unused bits must be zero. However, if a pattern was compiled with
401 PCRE_ANCHORED, or turned out to be anchored by virtue of its contents, it
402 cannot be made unachored at matching time.
403
404 There are also three further options that can be set only at matching time:
405
406 PCRE_NOTBOL
407
408 The first character of the string is not the beginning of a line, so the
409 circumflex metacharacter should not match before it. Setting this without
410 PCRE_MULTILINE (at compile time) causes circumflex never to match.
411
412 PCRE_NOTEOL
413
414 The end of the string is not the end of a line, so the dollar metacharacter
415 should not match it nor (except in multiline mode) a newline immediately before
416 it. Setting this without PCRE_MULTILINE (at compile time) causes dollar never
417 to match.
418
419 PCRE_NOTEMPTY
420
421 An empty string is not considered to be a valid match if this option is set. If
422 there are alternatives in the pattern, they are tried. If all the alternatives
423 match the empty string, the entire match fails. For example, if the pattern
424
425 a?b?
426
427 is applied to a string not beginning with "a" or "b", it matches the empty
428 string at the start of the subject. With PCRE_NOTEMPTY set, this match is not
429 valid, so PCRE searches further into the string for occurrences of "a" or "b".
430
431 Perl has no direct equivalent of PCRE_NOTEMPTY, but it does make a special case
432 of a pattern match of the empty string within its \fBsplit()\fR function, and
433 when using the /g modifier. It is possible to emulate Perl's behaviour after
434 matching a null string by first trying the match again at the same offset with
435 PCRE_NOTEMPTY set, and then if that fails by advancing the starting offset (see
436 below) and trying an ordinary match again.
437
438 The subject string is passed as a pointer in \fIsubject\fR, a length in
439 \fIlength\fR, and a starting offset in \fIstartoffset\fR. Unlike the pattern
440 string, it may contain binary zero characters. When the starting offset is
441 zero, the search for a match starts at the beginning of the subject, and this
442 is by far the most common case.
443
444 A non-zero starting offset is useful when searching for another match in the
445 same subject by calling \fBpcre_exec()\fR again after a previous success.
446 Setting \fIstartoffset\fR differs from just passing over a shortened string and
447 setting PCRE_NOTBOL in the case of a pattern that begins with any kind of
448 lookbehind. For example, consider the pattern
449
450 \\Biss\\B
451
452 which finds occurrences of "iss" in the middle of words. (\\B matches only if
453 the current position in the subject is not a word boundary.) When applied to
454 the string "Mississipi" the first call to \fBpcre_exec()\fR finds the first
455 occurrence. If \fBpcre_exec()\fR is called again with just the remainder of the
456 subject, namely "issipi", it does not match, because \\B is always false at the
457 start of the subject, which is deemed to be a word boundary. However, if
458 \fBpcre_exec()\fR is passed the entire string again, but with \fIstartoffset\fR
459 set to 4, it finds the second occurrence of "iss" because it is able to look
460 behind the starting point to discover that it is preceded by a letter.
461
462 If a non-zero starting offset is passed when the pattern is anchored, one
463 attempt to match at the given offset is tried. This can only succeed if the
464 pattern does not require the match to be at the start of the subject.
465
466 In general, a pattern matches a certain portion of the subject, and in
467 addition, further substrings from the subject may be picked out by parts of the
468 pattern. Following the usage in Jeffrey Friedl's book, this is called
469 "capturing" in what follows, and the phrase "capturing subpattern" is used for
470 a fragment of a pattern that picks out a substring. PCRE supports several other
471 kinds of parenthesized subpattern that do not cause substrings to be captured.
472
473 Captured substrings are returned to the caller via a vector of integer offsets
474 whose address is passed in \fIovector\fR. The number of elements in the vector
475 is passed in \fIovecsize\fR. The first two-thirds of the vector is used to pass
476 back captured substrings, each substring using a pair of integers. The
477 remaining third of the vector is used as workspace by \fBpcre_exec()\fR while
478 matching capturing subpatterns, and is not available for passing back
479 information. The length passed in \fIovecsize\fR should always be a multiple of
480 three. If it is not, it is rounded down.
481
482 When a match has been successful, information about captured substrings is
483 returned in pairs of integers, starting at the beginning of \fIovector\fR, and
484 continuing up to two-thirds of its length at the most. The first element of a
485 pair is set to the offset of the first character in a substring, and the second
486 is set to the offset of the first character after the end of a substring. The
487 first pair, \fIovector[0]\fR and \fIovector[1]\fR, identify the portion of the
488 subject string matched by the entire pattern. The next pair is used for the
489 first capturing subpattern, and so on. The value returned by \fBpcre_exec()\fR
490 is the number of pairs that have been set. If there are no capturing
491 subpatterns, the return value from a successful match is 1, indicating that
492 just the first pair of offsets has been set.
493
494 Some convenience functions are provided for extracting the captured substrings
495 as separate strings. These are described in the following section.
496
497 It is possible for an capturing subpattern number \fIn+1\fR to match some
498 part of the subject when subpattern \fIn\fR has not been used at all. For
499 example, if the string "abc" is matched against the pattern (a|(z))(bc)
500 subpatterns 1 and 3 are matched, but 2 is not. When this happens, both offset
501 values corresponding to the unused subpattern are set to -1.
502
503 If a capturing subpattern is matched repeatedly, it is the last portion of the
504 string that it matched that gets returned.
505
506 If the vector is too small to hold all the captured substrings, it is used as
507 far as possible (up to two-thirds of its length), and the function returns a
508 value of zero. In particular, if the substring offsets are not of interest,
509 \fBpcre_exec()\fR may be called with \fIovector\fR passed as NULL and
510 \fIovecsize\fR as zero. However, if the pattern contains back references and
511 the \fIovector\fR isn't big enough to remember the related substrings, PCRE has
512 to get additional memory for use during matching. Thus it is usually advisable
513 to supply an \fIovector\fR.
514
515 Note that \fBpcre_info()\fR can be used to find out how many capturing
516 subpatterns there are in a compiled pattern. The smallest size for
517 \fIovector\fR that will allow for \fIn\fR captured substrings in addition to
518 the offsets of the substring matched by the whole pattern is (\fIn\fR+1)*3.
519
520 If \fBpcre_exec()\fR fails, it returns a negative number. The following are
521 defined in the header file:
522
523 PCRE_ERROR_NOMATCH (-1)
524
525 The subject string did not match the pattern.
526
527 PCRE_ERROR_NULL (-2)
528
529 Either \fIcode\fR or \fIsubject\fR was passed as NULL, or \fIovector\fR was
530 NULL and \fIovecsize\fR was not zero.
531
532 PCRE_ERROR_BADOPTION (-3)
533
534 An unrecognized bit was set in the \fIoptions\fR argument.
535
536 PCRE_ERROR_BADMAGIC (-4)
537
538 PCRE stores a 4-byte "magic number" at the start of the compiled code, to catch
539 the case when it is passed a junk pointer. This is the error it gives when the
540 magic number isn't present.
541
542 PCRE_ERROR_UNKNOWN_NODE (-5)
543
544 While running the pattern match, an unknown item was encountered in the
545 compiled pattern. This error could be caused by a bug in PCRE or by overwriting
546 of the compiled pattern.
547
548 PCRE_ERROR_NOMEMORY (-6)
549
550 If a pattern contains back references, but the \fIovector\fR that is passed to
551 \fBpcre_exec()\fR is not big enough to remember the referenced substrings, PCRE
552 gets a block of memory at the start of matching to use for this purpose. If the
553 call via \fBpcre_malloc()\fR fails, this error is given. The memory is freed at
554 the end of matching.
555
556
557 .SH EXTRACTING CAPTURED SUBSTRINGS
558 Captured substrings can be accessed directly by using the offsets returned by
559 \fBpcre_exec()\fR in \fIovector\fR. For convenience, the functions
560 \fBpcre_copy_substring()\fR, \fBpcre_get_substring()\fR, and
561 \fBpcre_get_substring_list()\fR are provided for extracting captured substrings
562 as new, separate, zero-terminated strings. A substring that contains a binary
563 zero is correctly extracted and has a further zero added on the end, but the
564 result does not, of course, function as a C string.
565
566 The first three arguments are the same for all three functions: \fIsubject\fR
567 is the subject string which has just been successfully matched, \fIovector\fR
568 is a pointer to the vector of integer offsets that was passed to
569 \fBpcre_exec()\fR, and \fIstringcount\fR is the number of substrings that
570 were captured by the match, including the substring that matched the entire
571 regular expression. This is the value returned by \fBpcre_exec\fR if it
572 is greater than zero. If \fBpcre_exec()\fR returned zero, indicating that it
573 ran out of space in \fIovector\fR, the value passed as \fIstringcount\fR should
574 be the size of the vector divided by three.
575
576 The functions \fBpcre_copy_substring()\fR and \fBpcre_get_substring()\fR
577 extract a single substring, whose number is given as \fIstringnumber\fR. A
578 value of zero extracts the substring that matched the entire pattern, while
579 higher values extract the captured substrings. For \fBpcre_copy_substring()\fR,
580 the string is placed in \fIbuffer\fR, whose length is given by
581 \fIbuffersize\fR, while for \fBpcre_get_substring()\fR a new block of memory is
582 obtained via \fBpcre_malloc\fR, and its address is returned via
583 \fIstringptr\fR. The yield of the function is the length of the string, not
584 including the terminating zero, or one of
585
586 PCRE_ERROR_NOMEMORY (-6)
587
588 The buffer was too small for \fBpcre_copy_substring()\fR, or the attempt to get
589 memory failed for \fBpcre_get_substring()\fR.
590
591 PCRE_ERROR_NOSUBSTRING (-7)
592
593 There is no substring whose number is \fIstringnumber\fR.
594
595 The \fBpcre_get_substring_list()\fR function extracts all available substrings
596 and builds a list of pointers to them. All this is done in a single block of
597 memory which is obtained via \fBpcre_malloc\fR. The address of the memory block
598 is returned via \fIlistptr\fR, which is also the start of the list of string
599 pointers. The end of the list is marked by a NULL pointer. The yield of the
600 function is zero if all went well, or
601
602 PCRE_ERROR_NOMEMORY (-6)
603
604 if the attempt to get the memory block failed.
605
606 When any of these functions encounter a substring that is unset, which can
607 happen when capturing subpattern number \fIn+1\fR matches some part of the
608 subject, but subpattern \fIn\fR has not been used at all, they return an empty
609 string. This can be distinguished from a genuine zero-length substring by
610 inspecting the appropriate offset in \fIovector\fR, which is negative for unset
611 substrings.
612
613 The two convenience functions \fBpcre_free_substring()\fR and
614 \fBpcre_free_substring_list()\fR can be used to free the memory returned by
615 a previous call of \fBpcre_get_substring()\fR or
616 \fBpcre_get_substring_list()\fR, respectively. They do nothing more than call
617 the function pointed to by \fBpcre_free\fR, which of course could be called
618 directly from a C program. However, PCRE is used in some situations where it is
619 linked via a special interface to another programming language which cannot use
620 \fBpcre_free\fR directly; it is for these cases that the functions are
621 provided.
622
623
624 .SH LIMITATIONS
625 There are some size limitations in PCRE but it is hoped that they will never in
626 practice be relevant.
627 The maximum length of a compiled pattern is 65539 (sic) bytes.
628 All values in repeating quantifiers must be less than 65536.
629 The maximum number of capturing subpatterns is 99.
630 The maximum number of all parenthesized subpatterns, including capturing
631 subpatterns, assertions, and other types of subpattern, is 200.
632
633 The maximum length of a subject string is the largest positive number that an
634 integer variable can hold. However, PCRE uses recursion to handle subpatterns
635 and indefinite repetition. This means that the available stack space may limit
636 the size of a subject string that can be processed by certain patterns.
637
638
639 .SH DIFFERENCES FROM PERL
640 The differences described here are with respect to Perl 5.005.
641
642 1. By default, a whitespace character is any character that the C library
643 function \fBisspace()\fR recognizes, though it is possible to compile PCRE with
644 alternative character type tables. Normally \fBisspace()\fR matches space,
645 formfeed, newline, carriage return, horizontal tab, and vertical tab. Perl 5
646 no longer includes vertical tab in its set of whitespace characters. The \\v
647 escape that was in the Perl documentation for a long time was never in fact
648 recognized. However, the character itself was treated as whitespace at least
649 up to 5.002. In 5.004 and 5.005 it does not match \\s.
650
651 2. PCRE does not allow repeat quantifiers on lookahead assertions. Perl permits
652 them, but they do not mean what you might think. For example, (?!a){3} does
653 not assert that the next three characters are not "a". It just asserts that the
654 next character is not "a" three times.
655
656 3. Capturing subpatterns that occur inside negative lookahead assertions are
657 counted, but their entries in the offsets vector are never set. Perl sets its
658 numerical variables from any such patterns that are matched before the
659 assertion fails to match something (thereby succeeding), but only if the
660 negative lookahead assertion contains just one branch.
661
662 4. Though binary zero characters are supported in the subject string, they are
663 not allowed in a pattern string because it is passed as a normal C string,
664 terminated by zero. The escape sequence "\\0" can be used in the pattern to
665 represent a binary zero.
666
667 5. The following Perl escape sequences are not supported: \\l, \\u, \\L, \\U,
668 \\E, \\Q. In fact these are implemented by Perl's general string-handling and
669 are not part of its pattern matching engine.
670
671 6. The Perl \\G assertion is not supported as it is not relevant to single
672 pattern matches.
673
674 7. Fairly obviously, PCRE does not support the (?{code}) and (?p{code})
675 constructions. However, there is some experimental support for recursive
676 patterns using the non-Perl item (?R).
677
678 8. There are at the time of writing some oddities in Perl 5.005_02 concerned
679 with the settings of captured strings when part of a pattern is repeated. For
680 example, matching "aba" against the pattern /^(a(b)?)+$/ sets $2 to the value
681 "b", but matching "aabbaa" against /^(aa(bb)?)+$/ leaves $2 unset. However, if
682 the pattern is changed to /^(aa(b(b))?)+$/ then $2 (and $3) are set.
683
684 In Perl 5.004 $2 is set in both cases, and that is also true of PCRE. If in the
685 future Perl changes to a consistent state that is different, PCRE may change to
686 follow.
687
688 9. Another as yet unresolved discrepancy is that in Perl 5.005_02 the pattern
689 /^(a)?(?(1)a|b)+$/ matches the string "a", whereas in PCRE it does not.
690 However, in both Perl and PCRE /^(a)?a/ matched against "a" leaves $1 unset.
691
692 10. PCRE provides some extensions to the Perl regular expression facilities:
693
694 (a) Although lookbehind assertions must match fixed length strings, each
695 alternative branch of a lookbehind assertion can match a different length of
696 string. Perl 5.005 requires them all to have the same length.
697
698 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $ meta-
699 character matches only at the very end of the string.
700
701 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no special
702 meaning is faulted.
703
704 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quantifiers is
705 inverted, that is, by default they are not greedy, but if followed by a
706 question mark they are.
707
708 (e) PCRE_ANCHORED can be used to force a pattern to be tried only at the start
709 of the subject.
710
711 (f) The PCRE_NOTBOL, PCRE_NOTEOL, and PCRE_NOTEMPTY options for
712 \fBpcre_exec()\fR have no Perl equivalents.
713
714 (g) The (?R) construct allows for recursive pattern matching (Perl 5.6 can do
715 this using the (?p{code}) construct, which PCRE cannot of course support.)
716
717
718 .SH REGULAR EXPRESSION DETAILS
719 The syntax and semantics of the regular expressions supported by PCRE are
720 described below. Regular expressions are also described in the Perl
721 documentation and in a number of other books, some of which have copious
722 examples. Jeffrey Friedl's "Mastering Regular Expressions", published by
723 O'Reilly (ISBN 1-56592-257), covers them in great detail.
724
725 The description here is intended as reference documentation. The basic
726 operation of PCRE is on strings of bytes. However, there is the beginnings of
727 some support for UTF-8 character strings. To use this support you must
728 configure PCRE to include it, and then call \fBpcre_compile()\fR with the
729 PCRE_UTF8 option. How this affects the pattern matching is described in the
730 final section of this document.
731
732 A regular expression is a pattern that is matched against a subject string from
733 left to right. Most characters stand for themselves in a pattern, and match the
734 corresponding characters in the subject. As a trivial example, the pattern
735
736 The quick brown fox
737
738 matches a portion of a subject string that is identical to itself. The power of
739 regular expressions comes from the ability to include alternatives and
740 repetitions in the pattern. These are encoded in the pattern by the use of
741 \fImeta-characters\fR, which do not stand for themselves but instead are
742 interpreted in some special way.
743
744 There are two different sets of meta-characters: those that are recognized
745 anywhere in the pattern except within square brackets, and those that are
746 recognized in square brackets. Outside square brackets, the meta-characters are
747 as follows:
748
749 \\ general escape character with several uses
750 ^ assert start of subject (or line, in multiline mode)
751 $ assert end of subject (or line, in multiline mode)
752 . match any character except newline (by default)
753 [ start character class definition
754 | start of alternative branch
755 ( start subpattern
756 ) end subpattern
757 ? extends the meaning of (
758 also 0 or 1 quantifier
759 also quantifier minimizer
760 * 0 or more quantifier
761 + 1 or more quantifier
762 { start min/max quantifier
763
764 Part of a pattern that is in square brackets is called a "character class". In
765 a character class the only meta-characters are:
766
767 \\ general escape character
768 ^ negate the class, but only if the first character
769 - indicates character range
770 ] terminates the character class
771
772 The following sections describe the use of each of the meta-characters.
773
774
775 .SH BACKSLASH
776 The backslash character has several uses. Firstly, if it is followed by a
777 non-alphameric character, it takes away any special meaning that character may
778 have. This use of backslash as an escape character applies both inside and
779 outside character classes.
780
781 For example, if you want to match a "*" character, you write "\\*" in the
782 pattern. This applies whether or not the following character would otherwise be
783 interpreted as a meta-character, so it is always safe to precede a
784 non-alphameric with "\\" to specify that it stands for itself. In particular,
785 if you want to match a backslash, you write "\\\\".
786
787 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
788 pattern (other than in a character class) and characters between a "#" outside
789 a character class and the next newline character are ignored. An escaping
790 backslash can be used to include a whitespace or "#" character as part of the
791 pattern.
792
793 A second use of backslash provides a way of encoding non-printing characters
794 in patterns in a visible manner. There is no restriction on the appearance of
795 non-printing characters, apart from the binary zero that terminates a pattern,
796 but when a pattern is being prepared by text editing, it is usually easier to
797 use one of the following escape sequences than the binary character it
798 represents:
799
800 \\a alarm, that is, the BEL character (hex 07)
801 \\cx "control-x", where x is any character
802 \\e escape (hex 1B)
803 \\f formfeed (hex 0C)
804 \\n newline (hex 0A)
805 \\r carriage return (hex 0D)
806 \\t tab (hex 09)
807 \\xhh character with hex code hh
808 \\ddd character with octal code ddd, or backreference
809
810 The precise effect of "\\cx" is as follows: if "x" is a lower case letter, it
811 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
812 Thus "\\cz" becomes hex 1A, but "\\c{" becomes hex 3B, while "\\c;" becomes hex
813 7B.
814
815 After "\\x", up to two hexadecimal digits are read (letters can be in upper or
816 lower case).
817
818 After "\\0" up to two further octal digits are read. In both cases, if there
819 are fewer than two digits, just those that are present are used. Thus the
820 sequence "\\0\\x\\07" specifies two binary zeros followed by a BEL character.
821 Make sure you supply two digits after the initial zero if the character that
822 follows is itself an octal digit.
823
824 The handling of a backslash followed by a digit other than 0 is complicated.
825 Outside a character class, PCRE reads it and any following digits as a decimal
826 number. If the number is less than 10, or if there have been at least that many
827 previous capturing left parentheses in the expression, the entire sequence is
828 taken as a \fIback reference\fR. A description of how this works is given
829 later, following the discussion of parenthesized subpatterns.
830
831 Inside a character class, or if the decimal number is greater than 9 and there
832 have not been that many capturing subpatterns, PCRE re-reads up to three octal
833 digits following the backslash, and generates a single byte from the least
834 significant 8 bits of the value. Any subsequent digits stand for themselves.
835 For example:
836
837 \\040 is another way of writing a space
838 \\40 is the same, provided there are fewer than 40
839 previous capturing subpatterns
840 \\7 is always a back reference
841 \\11 might be a back reference, or another way of
842 writing a tab
843 \\011 is always a tab
844 \\0113 is a tab followed by the character "3"
845 \\113 is the character with octal code 113 (since there
846 can be no more than 99 back references)
847 \\377 is a byte consisting entirely of 1 bits
848 \\81 is either a back reference, or a binary zero
849 followed by the two characters "8" and "1"
850
851 Note that octal values of 100 or greater must not be introduced by a leading
852 zero, because no more than three octal digits are ever read.
853
854 All the sequences that define a single byte value can be used both inside and
855 outside character classes. In addition, inside a character class, the sequence
856 "\\b" is interpreted as the backspace character (hex 08). Outside a character
857 class it has a different meaning (see below).
858
859 The third use of backslash is for specifying generic character types:
860
861 \\d any decimal digit
862 \\D any character that is not a decimal digit
863 \\s any whitespace character
864 \\S any character that is not a whitespace character
865 \\w any "word" character
866 \\W any "non-word" character
867
868 Each pair of escape sequences partitions the complete set of characters into
869 two disjoint sets. Any given character matches one, and only one, of each pair.
870
871 A "word" character is any letter or digit or the underscore character, that is,
872 any character which can be part of a Perl "word". The definition of letters and
873 digits is controlled by PCRE's character tables, and may vary if locale-
874 specific matching is taking place (see "Locale support" above). For example, in
875 the "fr" (French) locale, some character codes greater than 128 are used for
876 accented letters, and these are matched by \\w.
877
878 These character type sequences can appear both inside and outside character
879 classes. They each match one character of the appropriate type. If the current
880 matching point is at the end of the subject string, all of them fail, since
881 there is no character to match.
882
883 The fourth use of backslash is for certain simple assertions. An assertion
884 specifies a condition that has to be met at a particular point in a match,
885 without consuming any characters from the subject string. The use of
886 subpatterns for more complicated assertions is described below. The backslashed
887 assertions are
888
889 \\b word boundary
890 \\B not a word boundary
891 \\A start of subject (independent of multiline mode)
892 \\Z end of subject or newline at end (independent of multiline mode)
893 \\z end of subject (independent of multiline mode)
894
895 These assertions may not appear in character classes (but note that "\\b" has a
896 different meaning, namely the backspace character, inside a character class).
897
898 A word boundary is a position in the subject string where the current character
899 and the previous character do not both match \\w or \\W (i.e. one matches
900 \\w and the other matches \\W), or the start or end of the string if the
901 first or last character matches \\w, respectively.
902
903 The \\A, \\Z, and \\z assertions differ from the traditional circumflex and
904 dollar (described below) in that they only ever match at the very start and end
905 of the subject string, whatever options are set. They are not affected by the
906 PCRE_NOTBOL or PCRE_NOTEOL options. If the \fIstartoffset\fR argument of
907 \fBpcre_exec()\fR is non-zero, \\A can never match. The difference between \\Z
908 and \\z is that \\Z matches before a newline that is the last character of the
909 string as well as at the end of the string, whereas \\z matches only at the
910 end.
911
912
913 .SH CIRCUMFLEX AND DOLLAR
914 Outside a character class, in the default matching mode, the circumflex
915 character is an assertion which is true only if the current matching point is
916 at the start of the subject string. If the \fIstartoffset\fR argument of
917 \fBpcre_exec()\fR is non-zero, circumflex can never match. Inside a character
918 class, circumflex has an entirely different meaning (see below).
919
920 Circumflex need not be the first character of the pattern if a number of
921 alternatives are involved, but it should be the first thing in each alternative
922 in which it appears if the pattern is ever to match that branch. If all
923 possible alternatives start with a circumflex, that is, if the pattern is
924 constrained to match only at the start of the subject, it is said to be an
925 "anchored" pattern. (There are also other constructs that can cause a pattern
926 to be anchored.)
927
928 A dollar character is an assertion which is true only if the current matching
929 point is at the end of the subject string, or immediately before a newline
930 character that is the last character in the string (by default). Dollar need
931 not be the last character of the pattern if a number of alternatives are
932 involved, but it should be the last item in any branch in which it appears.
933 Dollar has no special meaning in a character class.
934
935 The meaning of dollar can be changed so that it matches only at the very end of
936 the string, by setting the PCRE_DOLLAR_ENDONLY option at compile or matching
937 time. This does not affect the \\Z assertion.
938
939 The meanings of the circumflex and dollar characters are changed if the
940 PCRE_MULTILINE option is set. When this is the case, they match immediately
941 after and immediately before an internal "\\n" character, respectively, in
942 addition to matching at the start and end of the subject string. For example,
943 the pattern /^abc$/ matches the subject string "def\\nabc" in multiline mode,
944 but not otherwise. Consequently, patterns that are anchored in single line mode
945 because all branches start with "^" are not anchored in multiline mode, and a
946 match for circumflex is possible when the \fIstartoffset\fR argument of
947 \fBpcre_exec()\fR is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
948 PCRE_MULTILINE is set.
949
950 Note that the sequences \\A, \\Z, and \\z can be used to match the start and
951 end of the subject in both modes, and if all branches of a pattern start with
952 \\A is it always anchored, whether PCRE_MULTILINE is set or not.
953
954
955 .SH FULL STOP (PERIOD, DOT)
956 Outside a character class, a dot in the pattern matches any one character in
957 the subject, including a non-printing character, but not (by default) newline.
958 If the PCRE_DOTALL option is set, dots match newlines as well. The handling of
959 dot is entirely independent of the handling of circumflex and dollar, the only
960 relationship being that they both involve newline characters. Dot has no
961 special meaning in a character class.
962
963
964 .SH SQUARE BRACKETS
965 An opening square bracket introduces a character class, terminated by a closing
966 square bracket. A closing square bracket on its own is not special. If a
967 closing square bracket is required as a member of the class, it should be the
968 first data character in the class (after an initial circumflex, if present) or
969 escaped with a backslash.
970
971 A character class matches a single character in the subject; the character must
972 be in the set of characters defined by the class, unless the first character in
973 the class is a circumflex, in which case the subject character must not be in
974 the set defined by the class. If a circumflex is actually required as a member
975 of the class, ensure it is not the first character, or escape it with a
976 backslash.
977
978 For example, the character class [aeiou] matches any lower case vowel, while
979 [^aeiou] matches any character that is not a lower case vowel. Note that a
980 circumflex is just a convenient notation for specifying the characters which
981 are in the class by enumerating those that are not. It is not an assertion: it
982 still consumes a character from the subject string, and fails if the current
983 pointer is at the end of the string.
984
985 When caseless matching is set, any letters in a class represent both their
986 upper case and lower case versions, so for example, a caseless [aeiou] matches
987 "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
988 caseful version would.
989
990 The newline character is never treated in any special way in character classes,
991 whatever the setting of the PCRE_DOTALL or PCRE_MULTILINE options is. A class
992 such as [^a] will always match a newline.
993
994 The minus (hyphen) character can be used to specify a range of characters in a
995 character class. For example, [d-m] matches any letter between d and m,
996 inclusive. If a minus character is required in a class, it must be escaped with
997 a backslash or appear in a position where it cannot be interpreted as
998 indicating a range, typically as the first or last character in the class.
999
1000 It is not possible to have the literal character "]" as the end character of a
1001 range. A pattern such as [W-]46] is interpreted as a class of two characters
1002 ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
1003 "-46]". However, if the "]" is escaped with a backslash it is interpreted as
1004 the end of range, so [W-\\]46] is interpreted as a single class containing a
1005 range followed by two separate characters. The octal or hexadecimal
1006 representation of "]" can also be used to end a range.
1007
1008 Ranges operate in ASCII collating sequence. They can also be used for
1009 characters specified numerically, for example [\\000-\\037]. If a range that
1010 includes letters is used when caseless matching is set, it matches the letters
1011 in either case. For example, [W-c] is equivalent to [][\\^_`wxyzabc], matched
1012 caselessly, and if character tables for the "fr" locale are in use,
1013 [\\xc8-\\xcb] matches accented E characters in both cases.
1014
1015 The character types \\d, \\D, \\s, \\S, \\w, and \\W may also appear in a
1016 character class, and add the characters that they match to the class. For
1017 example, [\\dABCDEF] matches any hexadecimal digit. A circumflex can
1018 conveniently be used with the upper case character types to specify a more
1019 restricted set of characters than the matching lower case type. For example,
1020 the class [^\\W_] matches any letter or digit, but not underscore.
1021
1022 All non-alphameric characters other than \\, -, ^ (at the start) and the
1023 terminating ] are non-special in character classes, but it does no harm if they
1024 are escaped.
1025
1026
1027 .SH POSIX CHARACTER CLASSES
1028 Perl 5.6 (not yet released at the time of writing) is going to support the
1029 POSIX notation for character classes, which uses names enclosed by [: and :]
1030 within the enclosing square brackets. PCRE supports this notation. For example,
1031
1032 [01[:alpha:]%]
1033
1034 matches "0", "1", any alphabetic character, or "%". The supported class names
1035 are
1036
1037 alnum letters and digits
1038 alpha letters
1039 ascii character codes 0 - 127
1040 cntrl control characters
1041 digit decimal digits (same as \\d)
1042 graph printing characters, excluding space
1043 lower lower case letters
1044 print printing characters, including space
1045 punct printing characters, excluding letters and digits
1046 space white space (same as \\s)
1047 upper upper case letters
1048 word "word" characters (same as \\w)
1049 xdigit hexadecimal digits
1050
1051 The names "ascii" and "word" are Perl extensions. Another Perl extension is
1052 negation, which is indicated by a ^ character after the colon. For example,
1053
1054 [12[:^digit:]]
1055
1056 matches "1", "2", or any non-digit. PCRE (and Perl) also recogize the POSIX
1057 syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
1058 supported, and an error is given if they are encountered.
1059
1060
1061 .SH VERTICAL BAR
1062 Vertical bar characters are used to separate alternative patterns. For example,
1063 the pattern
1064
1065 gilbert|sullivan
1066
1067 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
1068 and an empty alternative is permitted (matching the empty string).
1069 The matching process tries each alternative in turn, from left to right,
1070 and the first one that succeeds is used. If the alternatives are within a
1071 subpattern (defined below), "succeeds" means matching the rest of the main
1072 pattern as well as the alternative in the subpattern.
1073
1074
1075 .SH INTERNAL OPTION SETTING
1076 The settings of PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and PCRE_EXTENDED
1077 can be changed from within the pattern by a sequence of Perl option letters
1078 enclosed between "(?" and ")". The option letters are
1079
1080 i for PCRE_CASELESS
1081 m for PCRE_MULTILINE
1082 s for PCRE_DOTALL
1083 x for PCRE_EXTENDED
1084
1085 For example, (?im) sets caseless, multiline matching. It is also possible to
1086 unset these options by preceding the letter with a hyphen, and a combined
1087 setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
1088 PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
1089 permitted. If a letter appears both before and after the hyphen, the option is
1090 unset.
1091
1092 The scope of these option changes depends on where in the pattern the setting
1093 occurs. For settings that are outside any subpattern (defined below), the
1094 effect is the same as if the options were set or unset at the start of
1095 matching. The following patterns all behave in exactly the same way:
1096
1097 (?i)abc
1098 a(?i)bc
1099 ab(?i)c
1100 abc(?i)
1101
1102 which in turn is the same as compiling the pattern abc with PCRE_CASELESS set.
1103 In other words, such "top level" settings apply to the whole pattern (unless
1104 there are other changes inside subpatterns). If there is more than one setting
1105 of the same option at top level, the rightmost setting is used.
1106
1107 If an option change occurs inside a subpattern, the effect is different. This
1108 is a change of behaviour in Perl 5.005. An option change inside a subpattern
1109 affects only that part of the subpattern that follows it, so
1110
1111 (a(?i)b)c
1112
1113 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
1114 By this means, options can be made to have different settings in different
1115 parts of the pattern. Any changes made in one alternative do carry on
1116 into subsequent branches within the same subpattern. For example,
1117
1118 (a(?i)b|c)
1119
1120 matches "ab", "aB", "c", and "C", even though when matching "C" the first
1121 branch is abandoned before the option setting. This is because the effects of
1122 option settings happen at compile time. There would be some very weird
1123 behaviour otherwise.
1124
1125 The PCRE-specific options PCRE_UNGREEDY and PCRE_EXTRA can be changed in the
1126 same way as the Perl-compatible options by using the characters U and X
1127 respectively. The (?X) flag setting is special in that it must always occur
1128 earlier in the pattern than any of the additional features it turns on, even
1129 when it is at top level. It is best put at the start.
1130
1131
1132 .SH SUBPATTERNS
1133 Subpatterns are delimited by parentheses (round brackets), which can be nested.
1134 Marking part of a pattern as a subpattern does two things:
1135
1136 1. It localizes a set of alternatives. For example, the pattern
1137
1138 cat(aract|erpillar|)
1139
1140 matches one of the words "cat", "cataract", or "caterpillar". Without the
1141 parentheses, it would match "cataract", "erpillar" or the empty string.
1142
1143 2. It sets up the subpattern as a capturing subpattern (as defined above).
1144 When the whole pattern matches, that portion of the subject string that matched
1145 the subpattern is passed back to the caller via the \fIovector\fR argument of
1146 \fBpcre_exec()\fR. Opening parentheses are counted from left to right (starting
1147 from 1) to obtain the numbers of the capturing subpatterns.
1148
1149 For example, if the string "the red king" is matched against the pattern
1150
1151 the ((red|white) (king|queen))
1152
1153 the captured substrings are "red king", "red", and "king", and are numbered 1,
1154 2, and 3.
1155
1156 The fact that plain parentheses fulfil two functions is not always helpful.
1157 There are often times when a grouping subpattern is required without a
1158 capturing requirement. If an opening parenthesis is followed by "?:", the
1159 subpattern does not do any capturing, and is not counted when computing the
1160 number of any subsequent capturing subpatterns. For example, if the string "the
1161 white queen" is matched against the pattern
1162
1163 the ((?:red|white) (king|queen))
1164
1165 the captured substrings are "white queen" and "queen", and are numbered 1 and
1166 2. The maximum number of captured substrings is 99, and the maximum number of
1167 all subpatterns, both capturing and non-capturing, is 200.
1168
1169 As a convenient shorthand, if any option settings are required at the start of
1170 a non-capturing subpattern, the option letters may appear between the "?" and
1171 the ":". Thus the two patterns
1172
1173 (?i:saturday|sunday)
1174 (?:(?i)saturday|sunday)
1175
1176 match exactly the same set of strings. Because alternative branches are tried
1177 from left to right, and options are not reset until the end of the subpattern
1178 is reached, an option setting in one branch does affect subsequent branches, so
1179 the above patterns match "SUNDAY" as well as "Saturday".
1180
1181
1182 .SH REPETITION
1183 Repetition is specified by quantifiers, which can follow any of the following
1184 items:
1185
1186 a single character, possibly escaped
1187 the . metacharacter
1188 a character class
1189 a back reference (see next section)
1190 a parenthesized subpattern (unless it is an assertion - see below)
1191
1192 The general repetition quantifier specifies a minimum and maximum number of
1193 permitted matches, by giving the two numbers in curly brackets (braces),
1194 separated by a comma. The numbers must be less than 65536, and the first must
1195 be less than or equal to the second. For example:
1196
1197 z{2,4}
1198
1199 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
1200 character. If the second number is omitted, but the comma is present, there is
1201 no upper limit; if the second number and the comma are both omitted, the
1202 quantifier specifies an exact number of required matches. Thus
1203
1204 [aeiou]{3,}
1205
1206 matches at least 3 successive vowels, but may match many more, while
1207
1208 \\d{8}
1209
1210 matches exactly 8 digits. An opening curly bracket that appears in a position
1211 where a quantifier is not allowed, or one that does not match the syntax of a
1212 quantifier, is taken as a literal character. For example, {,6} is not a
1213 quantifier, but a literal string of four characters.
1214
1215 The quantifier {0} is permitted, causing the expression to behave as if the
1216 previous item and the quantifier were not present.
1217
1218 For convenience (and historical compatibility) the three most common
1219 quantifiers have single-character abbreviations:
1220
1221 * is equivalent to {0,}
1222 + is equivalent to {1,}
1223 ? is equivalent to {0,1}
1224
1225 It is possible to construct infinite loops by following a subpattern that can
1226 match no characters with a quantifier that has no upper limit, for example:
1227
1228 (a?)*
1229
1230 Earlier versions of Perl and PCRE used to give an error at compile time for
1231 such patterns. However, because there are cases where this can be useful, such
1232 patterns are now accepted, but if any repetition of the subpattern does in fact
1233 match no characters, the loop is forcibly broken.
1234
1235 By default, the quantifiers are "greedy", that is, they match as much as
1236 possible (up to the maximum number of permitted times), without causing the
1237 rest of the pattern to fail. The classic example of where this gives problems
1238 is in trying to match comments in C programs. These appear between the
1239 sequences /* and */ and within the sequence, individual * and / characters may
1240 appear. An attempt to match C comments by applying the pattern
1241
1242 /\\*.*\\*/
1243
1244 to the string
1245
1246 /* first command */ not comment /* second comment */
1247
1248 fails, because it matches the entire string owing to the greediness of the .*
1249 item.
1250
1251 However, if a quantifier is followed by a question mark, it ceases to be
1252 greedy, and instead matches the minimum number of times possible, so the
1253 pattern
1254
1255 /\\*.*?\\*/
1256
1257 does the right thing with the C comments. The meaning of the various
1258 quantifiers is not otherwise changed, just the preferred number of matches.
1259 Do not confuse this use of question mark with its use as a quantifier in its
1260 own right. Because it has two uses, it can sometimes appear doubled, as in
1261
1262 \\d??\\d
1263
1264 which matches one digit by preference, but can match two if that is the only
1265 way the rest of the pattern matches.
1266
1267 If the PCRE_UNGREEDY option is set (an option which is not available in Perl),
1268 the quantifiers are not greedy by default, but individual ones can be made
1269 greedy by following them with a question mark. In other words, it inverts the
1270 default behaviour.
1271
1272 When a parenthesized subpattern is quantified with a minimum repeat count that
1273 is greater than 1 or with a limited maximum, more store is required for the
1274 compiled pattern, in proportion to the size of the minimum or maximum.
1275
1276 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
1277 to Perl's /s) is set, thus allowing the . to match newlines, the pattern is
1278 implicitly anchored, because whatever follows will be tried against every
1279 character position in the subject string, so there is no point in retrying the
1280 overall match at any position after the first. PCRE treats such a pattern as
1281 though it were preceded by \\A. In cases where it is known that the subject
1282 string contains no newlines, it is worth setting PCRE_DOTALL when the pattern
1283 begins with .* in order to obtain this optimization, or alternatively using ^
1284 to indicate anchoring explicitly.
1285
1286 When a capturing subpattern is repeated, the value captured is the substring
1287 that matched the final iteration. For example, after
1288
1289 (tweedle[dume]{3}\\s*)+
1290
1291 has matched "tweedledum tweedledee" the value of the captured substring is
1292 "tweedledee". However, if there are nested capturing subpatterns, the
1293 corresponding captured values may have been set in previous iterations. For
1294 example, after
1295
1296 /(a|(b))+/
1297
1298 matches "aba" the value of the second captured substring is "b".
1299
1300
1301 .SH BACK REFERENCES
1302 Outside a character class, a backslash followed by a digit greater than 0 (and
1303 possibly further digits) is a back reference to a capturing subpattern earlier
1304 (i.e. to its left) in the pattern, provided there have been that many previous
1305 capturing left parentheses.
1306
1307 However, if the decimal number following the backslash is less than 10, it is
1308 always taken as a back reference, and causes an error only if there are not
1309 that many capturing left parentheses in the entire pattern. In other words, the
1310 parentheses that are referenced need not be to the left of the reference for
1311 numbers less than 10. See the section entitled "Backslash" above for further
1312 details of the handling of digits following a backslash.
1313
1314 A back reference matches whatever actually matched the capturing subpattern in
1315 the current subject string, rather than anything matching the subpattern
1316 itself. So the pattern
1317
1318 (sens|respons)e and \\1ibility
1319
1320 matches "sense and sensibility" and "response and responsibility", but not
1321 "sense and responsibility". If caseful matching is in force at the time of the
1322 back reference, the case of letters is relevant. For example,
1323
1324 ((?i)rah)\\s+\\1
1325
1326 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1327 capturing subpattern is matched caselessly.
1328
1329 There may be more than one back reference to the same subpattern. If a
1330 subpattern has not actually been used in a particular match, any back
1331 references to it always fail. For example, the pattern
1332
1333 (a|(bc))\\2
1334
1335 always fails if it starts to match "a" rather than "bc". Because there may be
1336 up to 99 back references, all digits following the backslash are taken
1337 as part of a potential back reference number. If the pattern continues with a
1338 digit character, some delimiter must be used to terminate the back reference.
1339 If the PCRE_EXTENDED option is set, this can be whitespace. Otherwise an empty
1340 comment can be used.
1341
1342 A back reference that occurs inside the parentheses to which it refers fails
1343 when the subpattern is first used, so, for example, (a\\1) never matches.
1344 However, such references can be useful inside repeated subpatterns. For
1345 example, the pattern
1346
1347 (a|b\\1)+
1348
1349 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
1350 the subpattern, the back reference matches the character string corresponding
1351 to the previous iteration. In order for this to work, the pattern must be such
1352 that the first iteration does not need to match the back reference. This can be
1353 done using alternation, as in the example above, or by a quantifier with a
1354 minimum of zero.
1355
1356
1357 .SH ASSERTIONS
1358 An assertion is a test on the characters following or preceding the current
1359 matching point that does not actually consume any characters. The simple
1360 assertions coded as \\b, \\B, \\A, \\Z, \\z, ^ and $ are described above. More
1361 complicated assertions are coded as subpatterns. There are two kinds: those
1362 that look ahead of the current position in the subject string, and those that
1363 look behind it.
1364
1365 An assertion subpattern is matched in the normal way, except that it does not
1366 cause the current matching position to be changed. Lookahead assertions start
1367 with (?= for positive assertions and (?! for negative assertions. For example,
1368
1369 \\w+(?=;)
1370
1371 matches a word followed by a semicolon, but does not include the semicolon in
1372 the match, and
1373
1374 foo(?!bar)
1375
1376 matches any occurrence of "foo" that is not followed by "bar". Note that the
1377 apparently similar pattern
1378
1379 (?!foo)bar
1380
1381 does not find an occurrence of "bar" that is preceded by something other than
1382 "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1383 (?!foo) is always true when the next three characters are "bar". A
1384 lookbehind assertion is needed to achieve this effect.
1385
1386 Lookbehind assertions start with (?<= for positive assertions and (?<! for
1387 negative assertions. For example,
1388
1389 (?<!foo)bar
1390
1391 does find an occurrence of "bar" that is not preceded by "foo". The contents of
1392 a lookbehind assertion are restricted such that all the strings it matches must
1393 have a fixed length. However, if there are several alternatives, they do not
1394 all have to have the same fixed length. Thus
1395
1396 (?<=bullock|donkey)
1397
1398 is permitted, but
1399
1400 (?<!dogs?|cats?)
1401
1402 causes an error at compile time. Branches that match different length strings
1403 are permitted only at the top level of a lookbehind assertion. This is an
1404 extension compared with Perl 5.005, which requires all branches to match the
1405 same length of string. An assertion such as
1406
1407 (?<=ab(c|de))
1408
1409 is not permitted, because its single top-level branch can match two different
1410 lengths, but it is acceptable if rewritten to use two top-level branches:
1411
1412 (?<=abc|abde)
1413
1414 The implementation of lookbehind assertions is, for each alternative, to
1415 temporarily move the current position back by the fixed width and then try to
1416 match. If there are insufficient characters before the current position, the
1417 match is deemed to fail. Lookbehinds in conjunction with once-only subpatterns
1418 can be particularly useful for matching at the ends of strings; an example is
1419 given at the end of the section on once-only subpatterns.
1420
1421 Several assertions (of any sort) may occur in succession. For example,
1422
1423 (?<=\\d{3})(?<!999)foo
1424
1425 matches "foo" preceded by three digits that are not "999". Notice that each of
1426 the assertions is applied independently at the same point in the subject
1427 string. First there is a check that the previous three characters are all
1428 digits, and then there is a check that the same three characters are not "999".
1429 This pattern does \fInot\fR match "foo" preceded by six characters, the first
1430 of which are digits and the last three of which are not "999". For example, it
1431 doesn't match "123abcfoo". A pattern to do that is
1432
1433 (?<=\\d{3}...)(?<!999)foo
1434
1435 This time the first assertion looks at the preceding six characters, checking
1436 that the first three are digits, and then the second assertion checks that the
1437 preceding three characters are not "999".
1438
1439 Assertions can be nested in any combination. For example,
1440
1441 (?<=(?<!foo)bar)baz
1442
1443 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1444 preceded by "foo", while
1445
1446 (?<=\\d{3}(?!999)...)foo
1447
1448 is another pattern which matches "foo" preceded by three digits and any three
1449 characters that are not "999".
1450
1451 Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1452 because it makes no sense to assert the same thing several times. If any kind
1453 of assertion contains capturing subpatterns within it, these are counted for
1454 the purposes of numbering the capturing subpatterns in the whole pattern.
1455 However, substring capturing is carried out only for positive assertions,
1456 because it does not make sense for negative assertions.
1457
1458 Assertions count towards the maximum of 200 parenthesized subpatterns.
1459
1460
1461 .SH ONCE-ONLY SUBPATTERNS
1462 With both maximizing and minimizing repetition, failure of what follows
1463 normally causes the repeated item to be re-evaluated to see if a different
1464 number of repeats allows the rest of the pattern to match. Sometimes it is
1465 useful to prevent this, either to change the nature of the match, or to cause
1466 it fail earlier than it otherwise might, when the author of the pattern knows
1467 there is no point in carrying on.
1468
1469 Consider, for example, the pattern \\d+foo when applied to the subject line
1470
1471 123456bar
1472
1473 After matching all 6 digits and then failing to match "foo", the normal
1474 action of the matcher is to try again with only 5 digits matching the \\d+
1475 item, and then with 4, and so on, before ultimately failing. Once-only
1476 subpatterns provide the means for specifying that once a portion of the pattern
1477 has matched, it is not to be re-evaluated in this way, so the matcher would
1478 give up immediately on failing to match "foo" the first time. The notation is
1479 another kind of special parenthesis, starting with (?> as in this example:
1480
1481 (?>\\d+)bar
1482
1483 This kind of parenthesis "locks up" the part of the pattern it contains once
1484 it has matched, and a failure further into the pattern is prevented from
1485 backtracking into it. Backtracking past it to previous items, however, works as
1486 normal.
1487
1488 An alternative description is that a subpattern of this type matches the string
1489 of characters that an identical standalone pattern would match, if anchored at
1490 the current point in the subject string.
1491
1492 Once-only subpatterns are not capturing subpatterns. Simple cases such as the
1493 above example can be thought of as a maximizing repeat that must swallow
1494 everything it can. So, while both \\d+ and \\d+? are prepared to adjust the
1495 number of digits they match in order to make the rest of the pattern match,
1496 (?>\\d+) can only match an entire sequence of digits.
1497
1498 This construction can of course contain arbitrarily complicated subpatterns,
1499 and it can be nested.
1500
1501 Once-only subpatterns can be used in conjunction with lookbehind assertions to
1502 specify efficient matching at the end of the subject string. Consider a simple
1503 pattern such as
1504
1505 abcd$
1506
1507 when applied to a long string which does not match. Because matching proceeds
1508 from left to right, PCRE will look for each "a" in the subject and then see if
1509 what follows matches the rest of the pattern. If the pattern is specified as
1510
1511 ^.*abcd$
1512
1513 the initial .* matches the entire string at first, but when this fails (because
1514 there is no following "a"), it backtracks to match all but the last character,
1515 then all but the last two characters, and so on. Once again the search for "a"
1516 covers the entire string, from right to left, so we are no better off. However,
1517 if the pattern is written as
1518
1519 ^(?>.*)(?<=abcd)
1520
1521 there can be no backtracking for the .* item; it can match only the entire
1522 string. The subsequent lookbehind assertion does a single test on the last four
1523 characters. If it fails, the match fails immediately. For long strings, this
1524 approach makes a significant difference to the processing time.
1525
1526 When a pattern contains an unlimited repeat inside a subpattern that can itself
1527 be repeated an unlimited number of times, the use of a once-only subpattern is
1528 the only way to avoid some failing matches taking a very long time indeed.
1529 The pattern
1530
1531 (\\D+|<\\d+>)*[!?]
1532
1533 matches an unlimited number of substrings that either consist of non-digits, or
1534 digits enclosed in <>, followed by either ! or ?. When it matches, it runs
1535 quickly. However, if it is applied to
1536
1537 aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
1538
1539 it takes a long time before reporting failure. This is because the string can
1540 be divided between the two repeats in a large number of ways, and all have to
1541 be tried. (The example used [!?] rather than a single character at the end,
1542 because both PCRE and Perl have an optimization that allows for fast failure
1543 when a single character is used. They remember the last single character that
1544 is required for a match, and fail early if it is not present in the string.)
1545 If the pattern is changed to
1546
1547 ((?>\\D+)|<\\d+>)*[!?]
1548
1549 sequences of non-digits cannot be broken, and failure happens quickly.
1550
1551
1552 .SH CONDITIONAL SUBPATTERNS
1553 It is possible to cause the matching process to obey a subpattern
1554 conditionally or to choose between two alternative subpatterns, depending on
1555 the result of an assertion, or whether a previous capturing subpattern matched
1556 or not. The two possible forms of conditional subpattern are
1557
1558 (?(condition)yes-pattern)
1559 (?(condition)yes-pattern|no-pattern)
1560
1561 If the condition is satisfied, the yes-pattern is used; otherwise the
1562 no-pattern (if present) is used. If there are more than two alternatives in the
1563 subpattern, a compile-time error occurs.
1564
1565 There are two kinds of condition. If the text between the parentheses consists
1566 of a sequence of digits, the condition is satisfied if the capturing subpattern
1567 of that number has previously matched. The number must be greater than zero.
1568 Consider the following pattern, which contains non-significant white space to
1569 make it more readable (assume the PCRE_EXTENDED option) and to divide it into
1570 three parts for ease of discussion:
1571
1572 ( \\( )? [^()]+ (?(1) \\) )
1573
1574 The first part matches an optional opening parenthesis, and if that
1575 character is present, sets it as the first captured substring. The second part
1576 matches one or more characters that are not parentheses. The third part is a
1577 conditional subpattern that tests whether the first set of parentheses matched
1578 or not. If they did, that is, if subject started with an opening parenthesis,
1579 the condition is true, and so the yes-pattern is executed and a closing
1580 parenthesis is required. Otherwise, since no-pattern is not present, the
1581 subpattern matches nothing. In other words, this pattern matches a sequence of
1582 non-parentheses, optionally enclosed in parentheses.
1583
1584 If the condition is not a sequence of digits, it must be an assertion. This may
1585 be a positive or negative lookahead or lookbehind assertion. Consider this
1586 pattern, again containing non-significant white space, and with the two
1587 alternatives on the second line:
1588
1589 (?(?=[^a-z]*[a-z])
1590 \\d{2}-[a-z]{3}-\\d{2} | \\d{2}-\\d{2}-\\d{2} )
1591
1592 The condition is a positive lookahead assertion that matches an optional
1593 sequence of non-letters followed by a letter. In other words, it tests for the
1594 presence of at least one letter in the subject. If a letter is found, the
1595 subject is matched against the first alternative; otherwise it is matched
1596 against the second. This pattern matches strings in one of the two forms
1597 dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
1598
1599
1600 .SH COMMENTS
1601 The sequence (?# marks the start of a comment which continues up to the next
1602 closing parenthesis. Nested parentheses are not permitted. The characters
1603 that make up a comment play no part in the pattern matching at all.
1604
1605 If the PCRE_EXTENDED option is set, an unescaped # character outside a
1606 character class introduces a comment that continues up to the next newline
1607 character in the pattern.
1608
1609
1610 .SH RECURSIVE PATTERNS
1611 Consider the problem of matching a string in parentheses, allowing for
1612 unlimited nested parentheses. Without the use of recursion, the best that can
1613 be done is to use a pattern that matches up to some fixed depth of nesting. It
1614 is not possible to handle an arbitrary nesting depth. Perl 5.6 has provided an
1615 experimental facility that allows regular expressions to recurse (amongst other
1616 things). It does this by interpolating Perl code in the expression at run time,
1617 and the code can refer to the expression itself. A Perl pattern to solve the
1618 parentheses problem can be created like this:
1619
1620 $re = qr{\\( (?: (?>[^()]+) | (?p{$re}) )* \\)}x;
1621
1622 The (?p{...}) item interpolates Perl code at run time, and in this case refers
1623 recursively to the pattern in which it appears. Obviously, PCRE cannot support
1624 the interpolation of Perl code. Instead, the special item (?R) is provided for
1625 the specific case of recursion. This PCRE pattern solves the parentheses
1626 problem (assume the PCRE_EXTENDED option is set so that white space is
1627 ignored):
1628
1629 \\( ( (?>[^()]+) | (?R) )* \\)
1630
1631 First it matches an opening parenthesis. Then it matches any number of
1632 substrings which can either be a sequence of non-parentheses, or a recursive
1633 match of the pattern itself (i.e. a correctly parenthesized substring). Finally
1634 there is a closing parenthesis.
1635
1636 This particular example pattern contains nested unlimited repeats, and so the
1637 use of a once-only subpattern for matching strings of non-parentheses is
1638 important when applying the pattern to strings that do not match. For example,
1639 when it is applied to
1640
1641 (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
1642
1643 it yields "no match" quickly. However, if a once-only subpattern is not used,
1644 the match runs for a very long time indeed because there are so many different
1645 ways the + and * repeats can carve up the subject, and all have to be tested
1646 before failure can be reported.
1647
1648 The values set for any capturing subpatterns are those from the outermost level
1649 of the recursion at which the subpattern value is set. If the pattern above is
1650 matched against
1651
1652 (ab(cd)ef)
1653
1654 the value for the capturing parentheses is "ef", which is the last value taken
1655 on at the top level. If additional parentheses are added, giving
1656
1657 \\( ( ( (?>[^()]+) | (?R) )* ) \\)
1658 ^ ^
1659 ^ ^
1660 the string they capture is "ab(cd)ef", the contents of the top level
1661 parentheses. If there are more than 15 capturing parentheses in a pattern, PCRE
1662 has to obtain extra memory to store data during a recursion, which it does by
1663 using \fBpcre_malloc\fR, freeing it via \fBpcre_free\fR afterwards. If no
1664 memory can be obtained, it saves data for the first 15 capturing parentheses
1665 only, as there is no way to give an out-of-memory error from within a
1666 recursion.
1667
1668
1669 .SH PERFORMANCE
1670 Certain items that may appear in patterns are more efficient than others. It is
1671 more efficient to use a character class like [aeiou] than a set of alternatives
1672 such as (a|e|i|o|u). In general, the simplest construction that provides the
1673 required behaviour is usually the most efficient. Jeffrey Friedl's book
1674 contains a lot of discussion about optimizing regular expressions for efficient
1675 performance.
1676
1677 When a pattern begins with .* and the PCRE_DOTALL option is set, the pattern is
1678 implicitly anchored by PCRE, since it can match only at the start of a subject
1679 string. However, if PCRE_DOTALL is not set, PCRE cannot make this optimization,
1680 because the . metacharacter does not then match a newline, and if the subject
1681 string contains newlines, the pattern may match from the character immediately
1682 following one of them instead of from the very start. For example, the pattern
1683
1684 (.*) second
1685
1686 matches the subject "first\\nand second" (where \\n stands for a newline
1687 character) with the first captured substring being "and". In order to do this,
1688 PCRE has to retry the match starting after every newline in the subject.
1689
1690 If you are using such a pattern with subject strings that do not contain
1691 newlines, the best performance is obtained by setting PCRE_DOTALL, or starting
1692 the pattern with ^.* to indicate explicit anchoring. That saves PCRE from
1693 having to scan along the subject looking for a newline to restart at.
1694
1695 Beware of patterns that contain nested indefinite repeats. These can take a
1696 long time to run when applied to a string that does not match. Consider the
1697 pattern fragment
1698
1699 (a+)*
1700
1701 This can match "aaaa" in 33 different ways, and this number increases very
1702 rapidly as the string gets longer. (The * repeat can match 0, 1, 2, 3, or 4
1703 times, and for each of those cases other than 0, the + repeats can match
1704 different numbers of times.) When the remainder of the pattern is such that the
1705 entire match is going to fail, PCRE has in principle to try every possible
1706 variation, and this can take an extremely long time.
1707
1708 An optimization catches some of the more simple cases such as
1709
1710 (a+)*b
1711
1712 where a literal character follows. Before embarking on the standard matching
1713 procedure, PCRE checks that there is a "b" later in the subject string, and if
1714 there is not, it fails the match immediately. However, when there is no
1715 following literal this optimization cannot be used. You can see the difference
1716 by comparing the behaviour of
1717
1718 (a+)*\\d
1719
1720 with the pattern above. The former gives a failure almost instantly when
1721 applied to a whole line of "a" characters, whereas the latter takes an
1722 appreciable time with strings longer than about 20 characters.
1723
1724
1725 .SH UTF-8 SUPPORT
1726 Starting at release 3.3, PCRE has some support for character strings encoded
1727 in the UTF-8 format. This is incomplete, and is regarded as experimental. In
1728 order to use it, you must configure PCRE to include UTF-8 support in the code,
1729 and, in addition, you must call \fBpcre_compile()\fR with the PCRE_UTF8 option
1730 flag. When you do this, both the pattern and any subject strings that are
1731 matched against it are treated as UTF-8 strings instead of just strings of
1732 bytes, but only in the cases that are mentioned below.
1733
1734 If you compile PCRE with UTF-8 support, but do not use it at run time, the
1735 library will be a bit bigger, but the additional run time overhead is limited
1736 to testing the PCRE_UTF8 flag in several places, so should not be very large.
1737
1738 PCRE assumes that the strings it is given contain valid UTF-8 codes. It does
1739 not diagnose invalid UTF-8 strings. If you pass invalid UTF-8 strings to PCRE,
1740 the results are undefined.
1741
1742 Running with PCRE_UTF8 set causes these changes in the way PCRE works:
1743
1744 1. In a pattern, the escape sequence \\x{...}, where the contents of the braces
1745 is a string of hexadecimal digits, is interpreted as a UTF-8 character whose
1746 code number is the given hexadecimal number, for example: \\x{1234}. This
1747 inserts from one to six literal bytes into the pattern, using the UTF-8
1748 encoding. If a non-hexadecimal digit appears between the braces, the item is
1749 not recognized.
1750
1751 2. The original hexadecimal escape sequence, \\xhh, generates a two-byte UTF-8
1752 character if its value is greater than 127.
1753
1754 3. Repeat quantifiers are NOT correctly handled if they follow a multibyte
1755 character. For example, \\x{100}* and \\xc3+ do not work. If you want to
1756 repeat such characters, you must enclose them in non-capturing parentheses,
1757 for example (?:\\x{100}), at present.
1758
1759 4. The dot metacharacter matches one UTF-8 character instead of a single byte.
1760
1761 5. Unlike literal UTF-8 characters, the dot metacharacter followed by a
1762 repeat quantifier does operate correctly on UTF-8 characters instead of
1763 single bytes.
1764
1765 4. Although the \\x{...} escape is permitted in a character class, characters
1766 whose values are greater than 255 cannot be included in a class.
1767
1768 5. A class is matched against a UTF-8 character instead of just a single byte,
1769 but it can match only characters whose values are less than 256. Characters
1770 with greater values always fail to match a class.
1771
1772 6. Repeated classes work correctly on multiple characters.
1773
1774 7. Classes containing just a single character whose value is greater than 127
1775 (but less than 256), for example, [\\x80] or [^\\x{93}], do not work because
1776 these are optimized into single byte matches. In the first case, of course,
1777 the class brackets are just redundant.
1778
1779 8. Lookbehind assertions move backwards in the subject by a fixed number of
1780 characters instead of a fixed number of bytes. Simple cases have been tested
1781 to work correctly, but there may be hidden gotchas herein.
1782
1783 9. The character types such as \\d and \\w do not work correctly with UTF-8
1784 characters. They continue to test a single byte.
1785
1786 10. Anything not explicitly mentioned here continues to work in bytes rather
1787 than in characters.
1788
1789 The following UTF-8 features of Perl 5.6 are not implemented:
1790
1791 1. The escape sequence \\C to match a single byte.
1792
1793 2. The use of Unicode tables and properties and escapes \\p, \\P, and \\X.
1794
1795 .SH AUTHOR
1796 Philip Hazel <ph10@cam.ac.uk>
1797 .br
1798 University Computing Service,
1799 .br
1800 New Museums Site,
1801 .br
1802 Cambridge CB2 3QG, England.
1803 .br
1804 Phone: +44 1223 334714
1805
1806 Last updated: 28 August 2000,
1807 .br
1808 the 250th anniversary of the death of J.S. Bach.
1809 .br
1810 Copyright (c) 1997-2000 University of Cambridge.

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12