ViewVC logotype

Contents of /code/trunk/pcre.3

Parent Directory Parent Directory | Revision Log Revision Log

Revision 29 - (hide annotations) (download)
Sat Feb 24 21:38:53 2007 UTC (8 years, 2 months ago) by nigel
File size: 62604 byte(s)
Load pcre-2.03 into code/trunk.

1 nigel 3 .TH PCRE 3
2     .SH NAME
3     pcre - Perl-compatible regular expressions.
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     .B pcre_extra *pcre_study(const pcre *\fIcode\fR, int \fIoptions\fR,
17     .ti +5n
18 nigel 11 .B const char **\fIerrptr\fR);
19 nigel 3 .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 \fIoptions\fR,
24     .ti +5n
25     .B int *\fIovector\fR, int \fIovecsize\fR);
26     .PP
27     .br
28 nigel 29 .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 const unsigned char *pcre_maketables(void);
48     .PP
49     .br
50 nigel 3 .B int pcre_info(const pcre *\fIcode\fR, int *\fIoptptr\fR, int
51     .B *\fIfirstcharptr\fR);
52     .PP
53     .br
54     .B char *pcre_version(void);
55     .PP
56     .br
57     .B void *(*pcre_malloc)(size_t);
58     .PP
59     .br
60     .B void (*pcre_free)(void *);
65     The PCRE library is a set of functions that implement regular expression
66     pattern matching using the same syntax and semantics as Perl 5, with just a few
67 nigel 23 differences (see below). The current implementation corresponds to Perl 5.005.
68 nigel 3
69     PCRE has its own native API, which is described in this man page. There is also
70     a set of wrapper functions that correspond to the POSIX API. See
71     \fBpcreposix (3)\fR.
73 nigel 29 The functions \fBpcre_compile()\fR, \fBpcre_study()\fR, and \fBpcre_exec()\fR
74     are used for compiling and matching regular expressions, while
75     \fBpcre_copy_substring()\fR, \fBpcre_get_substring()\fR, and
76     \fBpcre_get_substring_list()\fR are convenience functions for extracting
77     captured substrings from a matched subject string. The function
78     \fBpcre_maketables()\fR is used (optionally) to build a set of character tables
79     in the current locale for passing to \fBpcre_compile()\fR.
80 nigel 25
81     The function \fBpcre_info()\fR is used to find out information about a compiled
82 nigel 3 pattern, while the function \fBpcre_version()\fR returns a pointer to a string
83     containing the version of PCRE and its date of release.
85     The global variables \fBpcre_malloc\fR and \fBpcre_free\fR initially contain
86     the entry points of the standard \fBmalloc()\fR and \fBfree()\fR functions
87     respectively. PCRE calls the memory management functions via these variables,
88     so a calling program can replace them if it wishes to intercept the calls. This
89     should be done before calling any PCRE functions.
93     The PCRE functions can be used in multi-threading applications, with the
94 nigel 25 proviso that the memory management functions pointed to by \fBpcre_malloc\fR
95     and \fBpcre_free\fR are shared by all threads.
96 nigel 3
97     The compiled form of a regular expression is not altered during matching, so
98     the same compiled pattern can safely be used by several threads at once.
102     The function \fBpcre_compile()\fR is called to compile a pattern into an
103     internal form. The pattern is a C string terminated by a binary zero, and
104 nigel 25 is passed in the argument \fIpattern\fR. A pointer to a single block of memory
105     that is obtained via \fBpcre_malloc\fR is returned. This contains the
106     compiled code and related data. The \fBpcre\fR type is defined for this for
107     convenience, but in fact \fBpcre\fR is just a typedef for \fBvoid\fR, since the
108     contents of the block are not externally defined. It is up to the caller to
109     free the memory when it is no longer required.
110 nigel 3 .PP
111     The size of a compiled pattern is roughly proportional to the length of the
112     pattern string, except that each character class (other than those containing
113     just a single character, negated or not) requires 33 bytes, and repeat
114     quantifiers with a minimum greater than one or a bounded maximum cause the
115     relevant portions of the compiled pattern to be replicated.
116     .PP
117     The \fIoptions\fR argument contains independent bits that affect the
118 nigel 23 compilation. It should be zero if no options are required. Some of the options,
119     in particular, those that are compatible with Perl, can also be set and unset
120     from within the pattern (see the detailed description of regular expressions
121     below). For these options, the contents of the \fIoptions\fR argument specifies
122     their initial settings at the start of compilation and execution. The
123     PCRE_ANCHORED option can be set at the time of matching as well as at compile
124     time.
125 nigel 3 .PP
126     If \fIerrptr\fR is NULL, \fBpcre_compile()\fR returns NULL immediately.
127     Otherwise, if compilation of a pattern fails, \fBpcre_compile()\fR returns
128     NULL, and sets the variable pointed to by \fIerrptr\fR to point to a textual
129 nigel 25 error message. The offset from the start of the pattern to the character where
130     the error was discovered is placed in the variable pointed to by
131     \fIerroffset\fR, which must not be NULL. If it is, an immediate error is given.
132 nigel 3 .PP
133 nigel 25 If the final argument, \fItableptr\fR, is NULL, PCRE uses a default set of
134     character tables which are built when it is compiled, using the default C
135     locale. Otherwise, \fItableptr\fR must be the result of a call to
136     \fBpcre_maketables()\fR. See the section on locale support below.
137     .PP
138 nigel 3 The following option bits are defined in the header file:
142     If this bit is set, the pattern is forced to be "anchored", that is, it is
143     constrained to match only at the start of the string which is being searched
144     (the "subject string"). This effect can also be achieved by appropriate
145     constructs in the pattern itself, which is the only way to do it in Perl.
149     If this bit is set, letters in the pattern match both upper and lower case
150 nigel 23 letters. It is equivalent to Perl's /i option.
151 nigel 3
154     If this bit is set, a dollar metacharacter in the pattern matches only at the
155 nigel 23 end of the subject string. Without this option, a dollar also matches
156     immediately before the final character if it is a newline (but not before any
157     other newlines). The PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is
158     set. There is no equivalent to this option in Perl.
159 nigel 3
162     If this bit is set, a dot metacharater in the pattern matches all characters,
163 nigel 23 including newlines. Without it, newlines are excluded. This option is
164 nigel 3 equivalent to Perl's /s option. A negative class such as [^a] always matches a
165     newline character, independent of the setting of this option.
169 nigel 23 If this bit is set, whitespace data characters in the pattern are totally
170     ignored except when escaped or inside a character class, and characters between
171     an unescaped # outside a character class and the next newline character,
172 nigel 3 inclusive, are also ignored. This is equivalent to Perl's /x option, and makes
173 nigel 23 it possible to include comments inside complicated patterns. Note, however,
174     that this applies only to data characters. Whitespace characters may never
175     appear within special character sequences in a pattern, for example within the
176     sequence (?( which introduces a conditional subpattern.
177 nigel 3
178 nigel 23 PCRE_EXTRA
180     This option turns on additional functionality of PCRE that is incompatible with
181     Perl. Any backslash in a pattern that is followed by a letter that has no
182     special meaning causes an error, thus reserving these combinations for future
183     expansion. By default, as in Perl, a backslash followed by a letter with no
184     special meaning is treated as a literal. There are at present no other features
185     controlled by this option.
187 nigel 3 PCRE_MULTILINE
189     By default, PCRE treats the subject string as consisting of a single "line" of
190     characters (even if it actually contains several newlines). The "start of line"
191     metacharacter (^) matches only at the start of the string, while the "end of
192     line" metacharacter ($) matches only at the end of the string, or before a
193 nigel 23 terminating newline (unless PCRE_DOLLAR_ENDONLY is set). This is the same as
194     Perl.
195 nigel 3
196     When PCRE_MULTILINE it is set, the "start of line" and "end of line" constructs
197     match immediately following or immediately before any newline in the subject
198     string, respectively, as well as at the very start and end. This is equivalent
199     to Perl's /m option. If there are no "\\n" characters in a subject string, or
200     no occurrences of ^ or $ in a pattern, setting PCRE_MULTILINE has no
201     effect.
203 nigel 19 PCRE_UNGREEDY
204 nigel 3
205 nigel 19 This option inverts the "greediness" of the quantifiers so that they are not
206     greedy by default, but become greedy if followed by "?". It is not compatible
207     with Perl. It can also be set by a (?U) option setting within the pattern.
208 nigel 3
209 nigel 19
211     When a pattern is going to be used several times, it is worth spending more
212     time analyzing it in order to speed up the time taken for matching. The
213     function \fBpcre_study()\fR takes a pointer to a compiled pattern as its first
214     argument, and returns a pointer to a \fBpcre_extra\fR block (another \fBvoid\fR
215     typedef) containing additional information about the pattern; this can be
216     passed to \fBpcre_exec()\fR. If no additional information is available, NULL
217     is returned.
219 nigel 23 The second argument contains option bits. At present, no options are defined
220     for \fBpcre_study()\fR, and this argument should always be zero.
221 nigel 3
222     The third argument for \fBpcre_study()\fR is a pointer to an error message. If
223     studying succeeds (even if no data is returned), the variable it points to is
224     set to NULL. Otherwise it points to a textual error message.
226     At present, studying a pattern is useful only for non-anchored patterns that do
227     not have a single fixed starting character. A bitmap of possible starting
228     characters is created.
231 nigel 25 .SH LOCALE SUPPORT
232     PCRE handles caseless matching, and determines whether characters are letters,
233     digits, or whatever, by reference to a set of tables. The library contains a
234     default set of tables which is created in the default C locale when PCRE is
235     compiled. This is used when the final argument of \fBpcre_compile()\fR is NULL,
236     and is sufficient for many applications.
238     An alternative set of tables can, however, be supplied. Such tables are built
239     by calling the \fBpcre_maketables()\fR function, which has no arguments, in the
240     relevant locale. The result can then be passed to \fBpcre_compile()\ as often
241     as necessary. For example, to build and use tables that are appropriate for the
242     French locale (where accented characters with codes greater than 128 are
243     treated as letters), the following code could be used:
245     setlocale(LC_CTYPE, "fr");
246     tables = pcre_maketables();
247     re = pcre_compile(..., tables);
249     The tables are built in memory that is obtained via \fBpcre_malloc\fR. The
250     pointer that is passed to \fBpcre_compile\fR is saved with the compiled
251     pattern, and the same tables are used via this pointer by \fBpcre_study()\fR
252     and \fBpcre_match()\fR. Thus for any single pattern, compilation, studying and
253     matching all happen in the same locale, but different patterns can be compiled
254     in different locales. It is the caller's responsibility to ensure that the
255     memory containing the tables remains available for as long as it is needed.
259     The \fBpcre_info()\fR function returns information about a compiled pattern.
260     Its yield is the number of capturing subpatterns, or one of the following
261     negative numbers:
263     PCRE_ERROR_NULL the argument \fIcode\fR was NULL
264     PCRE_ERROR_BADMAGIC the "magic number" was not found
266     If the \fIoptptr\fR argument is not NULL, a copy of the options with which the
267     pattern was compiled is placed in the integer it points to.
269     If the \fIfirstcharptr\fR argument is not NULL, is is used to pass back
270     information about the first character of any matched string. If there is a
271     fixed first character, e.g. from a pattern such as (cat|cow|coyote), then it is
272     returned in the integer pointed to by \fIfirstcharptr\fR. Otherwise, if the
273     pattern was compiled with the PCRE_MULTILINE option, and every branch started
274     with "^", then -1 is returned, indicating that the pattern will match at the
275     start of a subject string or after any "\\n" within the string. Otherwise -2 is
276     returned.
280     The function \fBpcre_exec()\fR is called to match a subject string against a
281     pre-compiled pattern, which is passed in the \fIcode\fR argument. If the
282     pattern has been studied, the result of the study should be passed in the
283     \fIextra\fR argument. Otherwise this must be NULL.
285     The subject string is passed as a pointer in \fIsubject\fR and a length in
286     \fIlength\fR. Unlike the pattern string, it may contain binary zero characters.
288 nigel 23 The PCRE_ANCHORED option can be passed in the \fIoptions\fR argument, whose
289     unused bits must be zero. However, if a pattern was compiled with
290     PCRE_ANCHORED, or turned out to be anchored by virtue of its contents, it
291     cannot be made unachored at matching time.
292 nigel 3
293     There are also two further options that can be set only at matching time:
297     The first character of the string is not the beginning of a line, so the
298     circumflex metacharacter should not match before it. Setting this without
299 nigel 23 PCRE_MULTILINE (at compile time) causes circumflex never to match.
300 nigel 3
303     The end of the string is not the end of a line, so the dollar metacharacter
304 nigel 23 should not match it nor (except in multiline mode) a newline immediately before
305     it. Setting this without PCRE_MULTILINE (at compile time) causes dollar never
306     to match.
307 nigel 3
308     In general, a pattern matches a certain portion of the subject, and in
309     addition, further substrings from the subject may be picked out by parts of the
310     pattern. Following the usage in Jeffrey Friedl's book, this is called
311     "capturing" in what follows, and the phrase "capturing subpattern" is used for
312     a fragment of a pattern that picks out a substring. PCRE supports several other
313     kinds of parenthesized subpattern that do not cause substrings to be captured.
315     Captured substrings are returned to the caller via a vector of integer offsets
316     whose address is passed in \fIovector\fR. The number of elements in the vector
317 nigel 23 is passed in \fIovecsize\fR. The first two-thirds of the vector is used to pass
318     back captured substrings, each substring using a pair of integers. The
319     remaining third of the vector is used as workspace by \fBpcre_exec()\fR while
320     matching capturing subpatterns, and is not available for passing back
321     information. The length passed in \fIovecsize\fR should always be a multiple of
322     three. If it is not, it is rounded down.
323 nigel 3
324 nigel 23 When a match has been successful, information about captured substrings is
325     returned in pairs of integers, starting at the beginning of \fIovector\fR, and
326     continuing up to two-thirds of its length at the most. The first element of a
327     pair is set to the offset of the first character in a substring, and the second
328     is set to the offset of the first character after the end of a substring. The
329     first pair, \fIovector[0]\fR and \fIovector[1]\fR, identify the portion of the
330     subject string matched by the entire pattern. The next pair is used for the
331     first capturing subpattern, and so on. The value returned by \fBpcre_exec()\fR
332     is the number of pairs that have been set. If there are no capturing
333     subpatterns, the return value from a successful match is 1, indicating that
334     just the first pair of offsets has been set.
335 nigel 3
336 nigel 29 Some convenience functions are provided for extracting the captured substrings
337     as separate strings. These are described in the following section.
339 nigel 3 It is possible for an capturing subpattern number \fIn+1\fR to match some
340     part of the subject when subpattern \fIn\fR has not been used at all. For
341 nigel 23 example, if the string "abc" is matched against the pattern (a|(z))(bc)
342 nigel 3 subpatterns 1 and 3 are matched, but 2 is not. When this happens, both offset
343     values corresponding to the unused subpattern are set to -1.
345     If a capturing subpattern is matched repeatedly, it is the last portion of the
346     string that it matched that gets returned.
348     If the vector is too small to hold all the captured substrings, it is used as
349 nigel 23 far as possible (up to two-thirds of its length), and the function returns a
350     value of zero. In particular, if the substring offsets are not of interest,
351     \fBpcre_exec()\fR may be called with \fIovector\fR passed as NULL and
352     \fIovecsize\fR as zero. However, if the pattern contains back references and
353     the \fIovector\fR isn't big enough to remember the related substrings, PCRE has
354     to get additional memory for use during matching. Thus it is usually advisable
355     to supply an \fIovector\fR.
356 nigel 3
357     Note that \fBpcre_info()\fR can be used to find out how many capturing
358 nigel 23 subpatterns there are in a compiled pattern. The smallest size for
359     \fIovector\fR that will allow for \fIn\fR captured substrings in addition to
360     the offsets of the substring matched by the whole pattern is (\fIn\fR+1)*3.
361 nigel 3
362     If \fBpcre_exec()\fR fails, it returns a negative number. The following are
363     defined in the header file:
367     The subject string did not match the pattern.
369 nigel 23 PCRE_ERROR_NULL (-2)
370 nigel 3
371     Either \fIcode\fR or \fIsubject\fR was passed as NULL, or \fIovector\fR was
372     NULL and \fIovecsize\fR was not zero.
374 nigel 23 PCRE_ERROR_BADOPTION (-3)
375 nigel 3
376     An unrecognized bit was set in the \fIoptions\fR argument.
378 nigel 23 PCRE_ERROR_BADMAGIC (-4)
379 nigel 3
380     PCRE stores a 4-byte "magic number" at the start of the compiled code, to catch
381     the case when it is passed a junk pointer. This is the error it gives when the
382     magic number isn't present.
384 nigel 23 PCRE_ERROR_UNKNOWN_NODE (-5)
385 nigel 3
386     While running the pattern match, an unknown item was encountered in the
387     compiled pattern. This error could be caused by a bug in PCRE or by overwriting
388     of the compiled pattern.
390 nigel 23 PCRE_ERROR_NOMEMORY (-6)
391 nigel 3
392     If a pattern contains back references, but the \fIovector\fR that is passed to
393     \fBpcre_exec()\fR is not big enough to remember the referenced substrings, PCRE
394     gets a block of memory at the start of matching to use for this purpose. If the
395     call via \fBpcre_malloc()\fR fails, this error is given. The memory is freed at
396     the end of matching.
400     Captured substrings can be accessed directly by using the offsets returned by
401     \fBpcre_exec()\fR in \fIovector\fR. For convenience, the functions
402     \fBpcre_copy_substring()\fR, \fBpcre_get_substring()\fR, and
403     \fBpcre_get_substring_list()\fR are provided for extracting captured substrings
404     as new, separate, zero-terminated strings. A substring that contains a binary
405     zero is correctly extracted and has a further zero added on the end, but the
406     result does not, of course, function as a C string.
407 nigel 3
408 nigel 29 The first three arguments are the same for all three functions: \fIsubject\fR
409     is the subject string which has just been successfully matched, \fIovector\fR
410     is a pointer to the vector of integer offsets that was passed to
411     \fBpcre_exec()\fR, and \fIstringcount\fR is the number of substrings that
412     were captured by the match, including the substring that matched the entire
413     regular expression. This is the value returned by \fBpcre_exec\fR if it
414     is greater than zero. If \fBpcre_exec()\fR returned zero, indicating that it
415     ran out of space in \fIovector\fR, then the value passed as
416     \fIstringcount\fR should be the size of the vector divided by three.
417 nigel 3
418 nigel 29 The functions \fBpcre_copy_substring()\fR and \fBpcre_get_substring()\fR
419     extract a single substring, whose number is given as \fIstringnumber\fR. A
420     value of zero extracts the substring that matched the entire pattern, while
421     higher values extract the captured substrings. For \fBpcre_copy_substring()\fR,
422     the string is placed in \fIbuffer\fR, whose length is given by
423     \fIbuffersize\fR, while for \fBpcre_get_substring()\fR a new block of store is
424     obtained via \fBpcre_malloc\fR, and its address is returned via
425     \fIstringptr\fR. The yield of the function is the length of the string, not
426     including the terminating zero, or one of
427 nigel 3
428 nigel 29 PCRE_ERROR_NOMEMORY (-6)
429 nigel 3
430 nigel 29 The buffer was too small for \fBpcre_copy_substring()\fR, or the attempt to get
431     memory failed for \fBpcre_get_substring()\fR.
432 nigel 3
433 nigel 29 PCRE_ERROR_NOSUBSTRING (-7)
435     There is no substring whose number is \fIstringnumber\fR.
437     The \fBpcre_get_substring_list()\fR function extracts all available substrings
438     and builds a list of pointers to them. All this is done in a single block of
439     memory which is obtained via \fBpcre_malloc\fR. The address of the memory block
440     is returned via \fIlistptr\fR, which is also the start of the list of string
441     pointers. The end of the list is marked by a NULL pointer. The yield of the
442     function is zero if all went well, or
446     if the attempt to get the memory block failed.
448     When any of these functions encounter a substring that is unset, which can
449     happen when capturing subpattern number \fIn+1\fR matches some part of the
450     subject, but subpattern \fIn\fR has not been used at all, they return an empty
451     string. This can be distinguished from a genuine zero-length substring by
452     inspecting the appropriate offset in \fIovector\fR, which is negative for unset
453     substrings.
457 nigel 3 .SH LIMITATIONS
458     There are some size limitations in PCRE but it is hoped that they will never in
459     practice be relevant.
460     The maximum length of a compiled pattern is 65539 (sic) bytes.
461     All values in repeating quantifiers must be less than 65536.
462     The maximum number of capturing subpatterns is 99.
463     The maximum number of all parenthesized subpatterns, including capturing
464 nigel 23 subpatterns, assertions, and other types of subpattern, is 200.
465 nigel 3
466     The maximum length of a subject string is the largest positive number that an
467     integer variable can hold. However, PCRE uses recursion to handle subpatterns
468     and indefinite repetition. This means that the available stack space may limit
469     the size of a subject string that can be processed by certain patterns.
473 nigel 23 The differences described here are with respect to Perl 5.005.
474 nigel 3
475     1. By default, a whitespace character is any character that the C library
476     function \fBisspace()\fR recognizes, though it is possible to compile PCRE with
477     alternative character type tables. Normally \fBisspace()\fR matches space,
478     formfeed, newline, carriage return, horizontal tab, and vertical tab. Perl 5
479     no longer includes vertical tab in its set of whitespace characters. The \\v
480     escape that was in the Perl documentation for a long time was never in fact
481     recognized. However, the character itself was treated as whitespace at least
482 nigel 23 up to 5.002. In 5.004 and 5.005 it does not match \\s.
483 nigel 3
484     2. PCRE does not allow repeat quantifiers on lookahead assertions. Perl permits
485 nigel 23 them, but they do not mean what you might think. For example, (?!a){3} does
486 nigel 3 not assert that the next three characters are not "a". It just asserts that the
487     next character is not "a" three times.
489     3. Capturing subpatterns that occur inside negative lookahead assertions are
490     counted, but their entries in the offsets vector are never set. Perl sets its
491     numerical variables from any such patterns that are matched before the
492     assertion fails to match something (thereby succeeding), but only if the
493     negative lookahead assertion contains just one branch.
495     4. Though binary zero characters are supported in the subject string, they are
496     not allowed in a pattern string because it is passed as a normal C string,
497     terminated by zero. The escape sequence "\\0" can be used in the pattern to
498     represent a binary zero.
500     5. The following Perl escape sequences are not supported: \\l, \\u, \\L, \\U,
501     \\E, \\Q. In fact these are implemented by Perl's general string-handling and
502     are not part of its pattern matching engine.
504     6. The Perl \\G assertion is not supported as it is not relevant to single
505     pattern matches.
507 nigel 23 7. Fairly obviously, PCRE does not support the (?{code}) construction.
508 nigel 3
509 nigel 23 8. There are at the time of writing some oddities in Perl 5.005_02 concerned
510     with the settings of captured strings when part of a pattern is repeated. For
511     example, matching "aba" against the pattern /^(a(b)?)+$/ sets $2 to the value
512     "b", but matching "aabbaa" against /^(aa(bb)?)+$/ leaves $2 unset. However, if
513     the pattern is changed to /^(aa(b(b))?)+$/ then $2 (and $3) get set.
514 nigel 3
515 nigel 23 In Perl 5.004 $2 is set in both cases, and that is also true of PCRE. If in the
516     future Perl changes to a consistent state that is different, PCRE may change to
517     follow.
518 nigel 3
519 nigel 23 9. Another as yet unresolved discrepancy is that in Perl 5.005_02 the pattern
520     /^(a)?(?(1)a|b)+$/ matches the string "a", whereas in PCRE it does not.
521     However, in both Perl and PCRE /^(a)?a/ matched against "a" leaves $1 unset.
522 nigel 3
523 nigel 23 10. PCRE provides some extensions to the Perl regular expression facilities:
524 nigel 3
525 nigel 23 (a) Although lookbehind assertions must match fixed length strings, each
526     alternative branch of a lookbehind assertion can match a different length of
527     string. Perl 5.005 requires them all to have the same length.
528 nigel 3
529 nigel 23 (b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $ meta-
530 nigel 3 character matches only at the very end of the string.
532 nigel 23 (c) If PCRE_EXTRA is set, a backslash followed by a letter with no special
533     meaning is faulted.
534 nigel 3
535 nigel 23 (d) If PCRE_UNGREEDY is set, the greediness of the repetition quantifiers is
536 nigel 19 inverted, that is, by default they are not greedy, but if followed by a
537     question mark they are.
538 nigel 3
539 nigel 19
541     The syntax and semantics of the regular expressions supported by PCRE are
542     described below. Regular expressions are also described in the Perl
543     documentation and in a number of other books, some of which have copious
544     examples. Jeffrey Friedl's "Mastering Regular Expressions", published by
545     O'Reilly (ISBN 1-56592-257-3), covers them in great detail. The description
546     here is intended as reference documentation.
548     A regular expression is a pattern that is matched against a subject string from
549     left to right. Most characters stand for themselves in a pattern, and match the
550     corresponding characters in the subject. As a trivial example, the pattern
552     The quick brown fox
554     matches a portion of a subject string that is identical to itself. The power of
555     regular expressions comes from the ability to include alternatives and
556     repetitions in the pattern. These are encoded in the pattern by the use of
557     \fImeta-characters\fR, which do not stand for themselves but instead are
558     interpreted in some special way.
560     There are two different sets of meta-characters: those that are recognized
561     anywhere in the pattern except within square brackets, and those that are
562     recognized in square brackets. Outside square brackets, the meta-characters are
563     as follows:
565     \\ general escape character with several uses
566     ^ assert start of subject (or line, in multiline mode)
567     $ assert end of subject (or line, in multiline mode)
568     . match any character except newline (by default)
569     [ start character class definition
570     | start of alternative branch
571     ( start subpattern
572     ) end subpattern
573     ? extends the meaning of (
574     also 0 or 1 quantifier
575     also quantifier minimizer
576     * 0 or more quantifier
577     + 1 or more quantifier
578     { start min/max quantifier
580     Part of a pattern that is in square brackets is called a "character class". In
581     a character class the only meta-characters are:
583     \\ general escape character
584     ^ negate the class, but only if the first character
585     - indicates character range
586     ] terminates the character class
588     The following sections describe the use of each of the meta-characters.
592     The backslash character has several uses. Firstly, if it is followed by a
593     non-alphameric character, it takes away any special meaning that character may
594     have. This use of backslash as an escape character applies both inside and
595     outside character classes.
597     For example, if you want to match a "*" character, you write "\\*" in the
598     pattern. This applies whether or not the following character would otherwise be
599     interpreted as a meta-character, so it is always safe to precede a
600     non-alphameric with "\\" to specify that it stands for itself. In particular,
601     if you want to match a backslash, you write "\\\\".
603     If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
604 nigel 23 pattern (other than in a character class) and characters between a "#" outside
605     a character class and the next newline character are ignored. An escaping
606     backslash can be used to include a whitespace or "#" character as part of the
607     pattern.
608 nigel 3
609     A second use of backslash provides a way of encoding non-printing characters
610     in patterns in a visible manner. There is no restriction on the appearance of
611     non-printing characters, apart from the binary zero that terminates a pattern,
612     but when a pattern is being prepared by text editing, it is usually easier to
613     use one of the following escape sequences than the binary character it
614     represents:
616     \\a alarm, that is, the BEL character (hex 07)
617     \\cx "control-x", where x is any character
618     \\e escape (hex 1B)
619     \\f formfeed (hex 0C)
620     \\n newline (hex 0A)
621     \\r carriage return (hex 0D)
622     \\t tab (hex 09)
623     \\xhh character with hex code hh
624 nigel 23 \\ddd character with octal code ddd, or backreference
625 nigel 3
626     The precise effect of "\\cx" is as follows: if "x" is a lower case letter, it
627     is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
628     Thus "\\cz" becomes hex 1A, but "\\c{" becomes hex 3B, while "\\c;" becomes hex
629     7B.
631     After "\\x", up to two hexadecimal digits are read (letters can be in upper or
632     lower case).
634     After "\\0" up to two further octal digits are read. In both cases, if there
635     are fewer than two digits, just those that are present are used. Thus the
636     sequence "\\0\\x\\07" specifies two binary zeros followed by a BEL character.
637 nigel 23 Make sure you supply two digits after the initial zero if the character that
638     follows is itself an octal digit.
639 nigel 3
640     The handling of a backslash followed by a digit other than 0 is complicated.
641     Outside a character class, PCRE reads it and any following digits as a decimal
642     number. If the number is less than 10, or if there have been at least that many
643     previous capturing left parentheses in the expression, the entire sequence is
644     taken as a \fIback reference\fR. A description of how this works is given
645     later, following the discussion of parenthesized subpatterns.
647     Inside a character class, or if the decimal number is greater than 9 and there
648     have not been that many capturing subpatterns, PCRE re-reads up to three octal
649     digits following the backslash, and generates a single byte from the least
650     significant 8 bits of the value. Any subsequent digits stand for themselves.
651     For example:
653     \\040 is another way of writing a space
654     \\40 is the same, provided there are fewer than 40
655     previous capturing subpatterns
656     \\7 is always a back reference
657     \\11 might be a back reference, or another way of
658     writing a tab
659     \\011 is always a tab
660     \\0113 is a tab followed by the character "3"
661     \\113 is the character with octal code 113 (since there
662     can be no more than 99 back references)
663     \\377 is a byte consisting entirely of 1 bits
664     \\81 is either a back reference, or a binary zero
665     followed by the two characters "8" and "1"
667     Note that octal values of 100 or greater must not be introduced by a leading
668     zero, because no more than three octal digits are ever read.
670     All the sequences that define a single byte value can be used both inside and
671     outside character classes. In addition, inside a character class, the sequence
672     "\\b" is interpreted as the backspace character (hex 08). Outside a character
673     class it has a different meaning (see below).
675     The third use of backslash is for specifying generic character types:
677     \\d any decimal digit
678     \\D any character that is not a decimal digit
679     \\s any whitespace character
680     \\S any character that is not a whitespace character
681     \\w any "word" character
682     \\W any "non-word" character
684     Each pair of escape sequences partitions the complete set of characters into
685     two disjoint sets. Any given character matches one, and only one, of each pair.
687     A "word" character is any letter or digit or the underscore character, that is,
688 nigel 25 any character which can be part of a Perl "word". The definition of letters and
689     digits is controlled by PCRE's character tables, and may vary if locale-
690     specific matching is taking place (see "Locale support" above). For example, in
691     the "fr" (French) locale, some character codes greater than 128 are used for
692     accented letters, and these are matched by \\w.
693 nigel 3
694 nigel 25 These character type sequences can appear both inside and outside character
695     classes. They each match one character of the appropriate type. If the current
696     matching point is at the end of the subject string, all of them fail, since
697     there is no character to match.
699 nigel 23 The fourth use of backslash is for certain simple assertions. An assertion
700     specifies a condition that has to be met at a particular point in a match,
701     without consuming any characters from the subject string. The use of
702     subpatterns for more complicated assertions is described below. The backslashed
703     assertions are
704 nigel 3
705     \\b word boundary
706     \\B not a word boundary
707     \\A start of subject (independent of multiline mode)
708 nigel 23 \\Z end of subject or newline at end (independent of multiline mode)
709     \\z end of subject (independent of multiline mode)
710 nigel 3
711 nigel 23 These assertions may not appear in character classes (but note that "\\b" has a
712 nigel 3 different meaning, namely the backspace character, inside a character class).
714     A word boundary is a position in the subject string where the current character
715 nigel 23 and the previous character do not both match \\w or \\W (i.e. one matches
716     \\w and the other matches \\W), or the start or end of the string if the
717     first or last character matches \\w, respectively.
718 nigel 3
719 nigel 23 The \\A, \\Z, and \\z assertions differ from the traditional circumflex and
720     dollar (described below) in that they only ever match at the very start and end
721     of the subject string, whatever options are set. They are not affected by the
722     PCRE_NOTBOL or PCRE_NOTEOL options. The difference between \\Z and \\z is that
723     \\Z matches before a newline that is the last character of the string as well
724     as at the end of the string, whereas \\z matches only at the end.
725 nigel 3
728 nigel 23 Outside a character class, in the default matching mode, the circumflex
729     character is an assertion which is true only if the current matching point is
730     at the start of the subject string. Inside a character class, circumflex has an
731 nigel 3 entirely different meaning (see below).
733     Circumflex need not be the first character of the pattern if a number of
734     alternatives are involved, but it should be the first thing in each alternative
735     in which it appears if the pattern is ever to match that branch. If all
736     possible alternatives start with a circumflex, that is, if the pattern is
737     constrained to match only at the start of the subject, it is said to be an
738     "anchored" pattern. (There are also other constructs that can cause a pattern
739     to be anchored.)
741     A dollar character is an assertion which is true only if the current matching
742     point is at the end of the subject string, or immediately before a newline
743     character that is the last character in the string (by default). Dollar need
744     not be the last character of the pattern if a number of alternatives are
745     involved, but it should be the last item in any branch in which it appears.
746     Dollar has no special meaning in a character class.
748     The meaning of dollar can be changed so that it matches only at the very end of
749     the string, by setting the PCRE_DOLLAR_ENDONLY option at compile or matching
750 nigel 23 time. This does not affect the \\Z assertion.
751 nigel 3
752     The meanings of the circumflex and dollar characters are changed if the
753 nigel 23 PCRE_MULTILINE option is set. When this is the case, they match immediately
754     after and immediately before an internal "\\n" character, respectively, in
755     addition to matching at the start and end of the subject string. For example,
756     the pattern /^abc$/ matches the subject string "def\\nabc" in multiline mode,
757     but not otherwise. Consequently, patterns that are anchored in single line mode
758     because all branches start with "^" are not anchored in multiline mode. The
759     PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
760 nigel 3
761 nigel 23 Note that the sequences \\A, \\Z, and \\z can be used to match the start and
762     end of the subject in both modes, and if all branches of a pattern start with
763     \\A is it always anchored, whether PCRE_MULTILINE is set or not.
764 nigel 3
767     Outside a character class, a dot in the pattern matches any one character in
768     the subject, including a non-printing character, but not (by default) newline.
769     If the PCRE_DOTALL option is set, then dots match newlines as well. The
770     handling of dot is entirely independent of the handling of circumflex and
771     dollar, the only relationship being that they both involve newline characters.
772     Dot has no special meaning in a character class.
776     An opening square bracket introduces a character class, terminated by a closing
777     square bracket. A closing square bracket on its own is not special. If a
778     closing square bracket is required as a member of the class, it should be the
779     first data character in the class (after an initial circumflex, if present) or
780 nigel 23 escaped with a backslash.
781 nigel 3
782     A character class matches a single character in the subject; the character must
783     be in the set of characters defined by the class, unless the first character in
784     the class is a circumflex, in which case the subject character must not be in
785     the set defined by the class. If a circumflex is actually required as a member
786 nigel 23 of the class, ensure it is not the first character, or escape it with a
787     backslash.
788 nigel 3
789     For example, the character class [aeiou] matches any lower case vowel, while
790     [^aeiou] matches any character that is not a lower case vowel. Note that a
791     circumflex is just a convenient notation for specifying the characters which
792     are in the class by enumerating those that are not. It is not an assertion: it
793     still consumes a character from the subject string, and fails if the current
794     pointer is at the end of the string.
796 nigel 25 When caseless matching is set, any letters in a class represent both their
797     upper case and lower case versions, so for example, a caseless [aeiou] matches
798     "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
799     caseful version would.
800 nigel 23
801 nigel 3 The newline character is never treated in any special way in character classes,
802     whatever the setting of the PCRE_DOTALL or PCRE_MULTILINE options is. A class
803     such as [^a] will always match a newline.
805     The minus (hyphen) character can be used to specify a range of characters in a
806     character class. For example, [d-m] matches any letter between d and m,
807     inclusive. If a minus character is required in a class, it must be escaped with
808 nigel 23 a backslash or appear in a position where it cannot be interpreted as
809 nigel 29 indicating a range, typically as the first or last character in the class.
810 nigel 3
811 nigel 29 It is not possible to have the literal character "]" as the end character of a
812     range. A pattern such as [W-]46] is interpreted as a class of two characters
813     ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
814     "-46]". However, if the "]" is escaped with a backslash it is interpreted as
815     the end of range, so [W-\\]46] is interpreted as a single class containing a
816     range followed by two separate characters. The octal or hexadecimal
817     representation of "]" can also be used to end a range.
819 nigel 3 Ranges operate in ASCII collating sequence. They can also be used for
820 nigel 25 characters specified numerically, for example [\\000-\\037]. If a range that
821     includes letters is used when caseless matching is set, it matches the letters
822     in either case. For example, [W-c] is equivalent to [][\\^_`wxyzabc], matched
823     caselessly, and if character tables for the "fr" locale are in use,
824     [\\xc8-\\xcb] matches accented E characters in both cases.
825 nigel 3
826     The character types \\d, \\D, \\s, \\S, \\w, and \\W may also appear in a
827     character class, and add the characters that they match to the class. For
828 nigel 23 example, [\\dABCDEF] matches any hexadecimal digit. A circumflex can
829     conveniently be used with the upper case character types to specify a more
830     restricted set of characters than the matching lower case type. For example,
831     the class [^\\W_] matches any letter or digit, but not underscore.
832 nigel 3
833     All non-alphameric characters other than \\, -, ^ (at the start) and the
834     terminating ] are non-special in character classes, but it does no harm if they
835     are escaped.
839 nigel 23 Vertical bar characters are used to separate alternative patterns. For example,
840     the pattern
841 nigel 3
842     gilbert|sullivan
844 nigel 23 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
845 nigel 3 and an empty alternative is permitted (matching the empty string).
846 nigel 23 The matching process tries each alternative in turn, from left to right,
847     and the first one that succeeds is used. If the alternatives are within a
848     subpattern (defined below), "succeeds" means matching the rest of the main
849     pattern as well as the alternative in the subpattern.
850 nigel 3
854     can be changed from within the pattern by a sequence of Perl option letters
855     enclosed between "(?" and ")". The option letters are
857     i for PCRE_CASELESS
858     m for PCRE_MULTILINE
859     s for PCRE_DOTALL
860     x for PCRE_EXTENDED
862     For example, (?im) sets caseless, multiline matching. It is also possible to
863     unset these options by preceding the letter with a hyphen, and a combined
864     setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
865     PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
866     permitted. If a letter appears both before and after the hyphen, the option is
867     unset.
869     The scope of these option changes depends on where in the pattern the setting
870     occurs. For settings that are outside any subpattern (defined below), the
871     effect is the same as if the options were set or unset at the start of
872     matching. The following patterns all behave in exactly the same way:
874     (?i)abc
875     a(?i)bc
876     ab(?i)c
877     abc(?i)
879     which in turn is the same as compiling the pattern abc with PCRE_CASELESS set.
880     In other words, such "top level" settings apply to the whole pattern (unless
881     there are other changes inside subpatterns). If there is more than one setting
882     of the same option at top level, the rightmost setting is used.
884     If an option change occurs inside a subpattern, the effect is different. This
885     is a change of behaviour in Perl 5.005. An option change inside a subpattern
886     affects only that part of the subpattern that follows it, so
888     (a(?i)b)c
890     matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
891     By this means, options can be made to have different settings in different
892     parts of the pattern. Any changes made in one alternative do carry on
893     into subsequent branches within the same subpattern. For example,
895     (a(?i)b|c)
897     matches "ab", "aB", "c", and "C", even though when matching "C" the first
898     branch is abandoned before the option setting. This is because the effects of
899     option settings happen at compile time. There would be some very weird
900     behaviour otherwise.
902     The PCRE-specific options PCRE_UNGREEDY and PCRE_EXTRA can be changed in the
903     same way as the Perl-compatible options by using the characters U and X
904     respectively. The (?X) flag setting is special in that it must always occur
905     earlier in the pattern than any of the additional features it turns on, even
906     when it is at top level. It is best put at the start.
909 nigel 3 .SH SUBPATTERNS
910     Subpatterns are delimited by parentheses (round brackets), which can be nested.
911     Marking part of a pattern as a subpattern does two things:
913     1. It localizes a set of alternatives. For example, the pattern
915     cat(aract|erpillar|)
917     matches one of the words "cat", "cataract", or "caterpillar". Without the
918     parentheses, it would match "cataract", "erpillar" or the empty string.
920     2. It sets up the subpattern as a capturing subpattern (as defined above).
921     When the whole pattern matches, that portion of the subject string that matched
922     the subpattern is passed back to the caller via the \fIovector\fR argument of
923     \fBpcre_exec()\fR. Opening parentheses are counted from left to right (starting
924     from 1) to obtain the numbers of the capturing subpatterns.
926     For example, if the string "the red king" is matched against the pattern
928     the ((red|white) (king|queen))
930     the captured substrings are "red king", "red", and "king", and are numbered 1,
931     2, and 3.
933     The fact that plain parentheses fulfil two functions is not always helpful.
934     There are often times when a grouping subpattern is required without a
935     capturing requirement. If an opening parenthesis is followed by "?:", the
936     subpattern does not do any capturing, and is not counted when computing the
937     number of any subsequent capturing subpatterns. For example, if the string "the
938     white queen" is matched against the pattern
940     the ((?:red|white) (king|queen))
942     the captured substrings are "white queen" and "queen", and are numbered 1 and
943     2. The maximum number of captured substrings is 99, and the maximum number of
944     all subpatterns, both capturing and non-capturing, is 200.
946 nigel 23 As a convenient shorthand, if any option settings are required at the start of
947     a non-capturing subpattern, the option letters may appear between the "?" and
948     the ":". Thus the two patterns
949 nigel 3
950 nigel 23 (?i:saturday|sunday)
951     (?:(?i)saturday|sunday)
952 nigel 3
953 nigel 23 match exactly the same set of strings. Because alternative branches are tried
954     from left to right, and options are not reset until the end of the subpattern
955     is reached, an option setting in one branch does affect subsequent branches, so
956     the above patterns match "SUNDAY" as well as "Saturday".
957 nigel 3
960     Repetition is specified by quantifiers, which can follow any of the following
961     items:
963     a single character, possibly escaped
964     the . metacharacter
965     a character class
966 nigel 23 a back reference (see next section)
967     a parenthesized subpattern (unless it is an assertion - see below)
968 nigel 3
969     The general repetition quantifier specifies a minimum and maximum number of
970     permitted matches, by giving the two numbers in curly brackets (braces),
971     separated by a comma. The numbers must be less than 65536, and the first must
972     be less than or equal to the second. For example:
974     z{2,4}
976     matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
977     character. If the second number is omitted, but the comma is present, there is
978     no upper limit; if the second number and the comma are both omitted, the
979     quantifier specifies an exact number of required matches. Thus
981     [aeiou]{3,}
983     matches at least 3 successive vowels, but may match many more, while
985     \\d{8}
987     matches exactly 8 digits. An opening curly bracket that appears in a position
988     where a quantifier is not allowed, or one that does not match the syntax of a
989 nigel 23 quantifier, is taken as a literal character. For example, {,6} is not a
990 nigel 3 quantifier, but a literal string of four characters.
992     The quantifier {0} is permitted, causing the expression to behave as if the
993     previous item and the quantifier were not present.
995     For convenience (and historical compatibility) the three most common
996     quantifiers have single-character abbreviations:
998     * is equivalent to {0,}
999     + is equivalent to {1,}
1000     ? is equivalent to {0,1}
1002 nigel 23 It is possible to construct infinite loops by following a subpattern that can
1003     match no characters with a quantifier that has no upper limit, for example:
1005     (a?)*
1007     Earlier versions of Perl and PCRE used to give an error at compile time for
1008     such patterns. However, because there are cases where this can be useful, such
1009     patterns are now accepted, but if any repetition of the subpattern does in fact
1010     match no characters, the loop is forcibly broken.
1012 nigel 3 By default, the quantifiers are "greedy", that is, they match as much as
1013     possible (up to the maximum number of permitted times), without causing the
1014     rest of the pattern to fail. The classic example of where this gives problems
1015     is in trying to match comments in C programs. These appear between the
1016     sequences /* and */ and within the sequence, individual * and / characters may
1017     appear. An attempt to match C comments by applying the pattern
1019     /\\*.*\\*/
1021     to the string
1023     /* first command */ not comment /* second comment */
1025     fails, because it matches the entire string due to the greediness of the .*
1026     item.
1028     However, if a quantifier is followed by a question mark, then it ceases to be
1029     greedy, and instead matches the minimum number of times possible, so the
1030     pattern
1032     /\\*.*?\\*/
1034     does the right thing with the C comments. The meaning of the various
1035     quantifiers is not otherwise changed, just the preferred number of matches.
1036     Do not confuse this use of question mark with its use as a quantifier in its
1037     own right. Because it has two uses, it can sometimes appear doubled, as in
1039 nigel 23 \\d??\\d
1040 nigel 3
1041     which matches one digit by preference, but can match two if that is the only
1042     way the rest of the pattern matches.
1044 nigel 19 If the PCRE_UNGREEDY option is set (an option which is not available in Perl)
1045     then the quantifiers are not greedy by default, but individual ones can be made
1046 nigel 23 greedy by following them with a question mark. In other words, it inverts the
1047 nigel 19 default behaviour.
1049 nigel 7 When a parenthesized subpattern is quantified with a minimum repeat count that
1050 nigel 3 is greater than 1 or with a limited maximum, more store is required for the
1051     compiled pattern, in proportion to the size of the minimum or maximum.
1053     If a pattern starts with .* then it is implicitly anchored, since whatever
1054     follows will be tried against every character position in the subject string.
1055     PCRE treats this as though it were preceded by \\A.
1057     When a capturing subpattern is repeated, the value captured is the substring
1058 nigel 23 that matched the final iteration. For example, after
1059 nigel 3
1060 nigel 23 (tweedle[dume]{3}\\s*)+
1061 nigel 3
1062 nigel 23 has matched "tweedledum tweedledee" the value of the captured substring is
1063     "tweedledee". However, if there are nested capturing subpatterns, the
1064     corresponding captured values may have been set in previous iterations. For
1065     example, after
1066 nigel 3
1067 nigel 23 /(a|(b))+/
1068 nigel 3
1069 nigel 23 matches "aba" the value of the second captured substring is "b".
1073     Outside a character class, a backslash followed by a digit greater than 0 (and
1074     possibly further digits) is a back reference to a capturing subpattern earlier
1075     (i.e. to its left) in the pattern, provided there have been that many previous
1076     capturing left parentheses.
1078     However, if the decimal number following the backslash is less than 10, it is
1079     always taken as a back reference, and causes an error only if there are not
1080     that many capturing left parentheses in the entire pattern. In other words, the
1081     parentheses that are referenced need not be to the left of the reference for
1082     numbers less than 10. See the section entitled "Backslash" above for further
1083     details of the handling of digits following a backslash.
1085     A back reference matches whatever actually matched the capturing subpattern in
1086     the current subject string, rather than anything matching the subpattern
1087     itself. So the pattern
1089     (sens|respons)e and \\1ibility
1091     matches "sense and sensibility" and "response and responsibility", but not
1092     "sense and responsibility". If caseful matching is in force at the time of the
1093     back reference, then the case of letters is relevant. For example,
1095     ((?i)rah)\\s+\\1
1097     matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
1098     capturing subpattern is matched caselessly.
1100     There may be more than one back reference to the same subpattern. If a
1101     subpattern has not actually been used in a particular match, then any back
1102     references to it always fail. For example, the pattern
1104     (a|(bc))\\2
1106     always fails if it starts to match "a" rather than "bc". Because there may be
1107     up to 99 back references, all digits following the backslash are taken
1108     as part of a potential back reference number. If the pattern continues with a
1109     digit character, then some delimiter must be used to terminate the back
1110     reference. If the PCRE_EXTENDED option is set, this can be whitespace.
1111     Otherwise an empty comment can be used.
1113     A back reference that occurs inside the parentheses to which it refers fails
1114     when the subpattern is first used, so, for example, (a\\1) never matches.
1115     However, such references can be useful inside repeated subpatterns. For
1116     example, the pattern
1118     (a|b\\1)+
1120     matches any number of "a"s and also "aba", "ababaa" etc. At each iteration of
1121     the subpattern, the back reference matches the character string corresponding
1122     to the previous iteration. In order for this to work, the pattern must be such
1123     that the first iteration does not need to match the back reference. This can be
1124     done using alternation, as in the example above, or by a quantifier with a
1125     minimum of zero.
1128 nigel 3 .SH ASSERTIONS
1129 nigel 23 An assertion is a test on the characters following or preceding the current
1130     matching point that does not actually consume any characters. The simple
1131     assertions coded as \\b, \\B, \\A, \\Z, \\z, ^ and $ are described above. More
1132     complicated assertions are coded as subpatterns. There are two kinds: those
1133     that look ahead of the current position in the subject string, and those that
1134     look behind it.
1135 nigel 3
1136 nigel 23 An assertion subpattern is matched in the normal way, except that it does not
1137     cause the current matching position to be changed. Lookahead assertions start
1138     with (?= for positive assertions and (?! for negative assertions. For example,
1140 nigel 3 \\w+(?=;)
1142     matches a word followed by a semicolon, but does not include the semicolon in
1143     the match, and
1145     foo(?!bar)
1147     matches any occurrence of "foo" that is not followed by "bar". Note that the
1148     apparently similar pattern
1150     (?!foo)bar
1152     does not find an occurrence of "bar" that is preceded by something other than
1153     "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
1154 nigel 23 (?!foo) is always true when the next three characters are "bar". A
1155     lookbehind assertion is needed to achieve this effect.
1156 nigel 3
1157 nigel 23 Lookbehind assertions start with (?<= for positive assertions and (?<! for
1158     negative assertions. For example,
1160     (?<!foo)bar
1162     does find an occurrence of "bar" that is not preceded by "foo". The contents of
1163     a lookbehind assertion are restricted such that all the strings it matches must
1164     have a fixed length. However, if there are several alternatives, they do not
1165     all have to have the same fixed length. Thus
1167     (?<=bullock|donkey)
1169     is permitted, but
1171     (?<!dogs?|cats?)
1173     causes an error at compile time. Branches that match different length strings
1174     are permitted only at the top level of a lookbehind assertion. This is an
1175     extension compared with Perl 5.005, which requires all branches to match the
1176     same length of string. An assertion such as
1178     (?<=ab(c|de))
1180 nigel 27 is not permitted, because its single top-level branch can match two different
1181     lengths, but it is acceptable if rewritten to use two top-level branches:
1182 nigel 23
1183     (?<=abc|abde)
1185     The implementation of lookbehind assertions is, for each alternative, to
1186     temporarily move the current position back by the fixed width and then try to
1187     match. If there are insufficient characters before the current position, the
1188 nigel 27 match is deemed to fail. Lookbehinds in conjunction with once-only subpatterns
1189     can be particularly useful for matching at the ends of strings; an example is
1190     given at the end of the section on once-only subpatterns.
1191 nigel 23
1192 nigel 27 Several assertions (of any sort) may occur in succession. For example,
1193 nigel 23
1194 nigel 27 (?<=\\d{3})(?<!999)foo
1196     matches "foo" preceded by three digits that are not "999". Furthermore,
1197     assertions can be nested in any combination. For example,
1199 nigel 23 (?<=(?<!foo)bar)baz
1201     matches an occurrence of "baz" that is preceded by "bar" which in turn is not
1202     preceded by "foo".
1204 nigel 3 Assertion subpatterns are not capturing subpatterns, and may not be repeated,
1205     because it makes no sense to assert the same thing several times. If an
1206     assertion contains capturing subpatterns within it, these are always counted
1207     for the purposes of numbering the capturing subpatterns in the whole pattern.
1208     Substring capturing is carried out for positive assertions, but it does not
1209     make sense for negative assertions.
1211     Assertions count towards the maximum of 200 parenthesized subpatterns.
1215     With both maximizing and minimizing repetition, failure of what follows
1216     normally causes the repeated item to be re-evaluated to see if a different
1217     number of repeats allows the rest of the pattern to match. Sometimes it is
1218     useful to prevent this, either to change the nature of the match, or to cause
1219 nigel 23 it fail earlier than it otherwise might, when the author of the pattern knows
1220 nigel 3 there is no point in carrying on.
1222     Consider, for example, the pattern \\d+foo when applied to the subject line
1224 nigel 23 123456bar
1225 nigel 3
1226     After matching all 6 digits and then failing to match "foo", the normal
1227     action of the matcher is to try again with only 5 digits matching the \\d+
1228     item, and then with 4, and so on, before ultimately failing. Once-only
1229     subpatterns provide the means for specifying that once a portion of the pattern
1230     has matched, it is not to be re-evaluated in this way, so the matcher would
1231     give up immediately on failing to match "foo" the first time. The notation is
1232     another kind of special parenthesis, starting with (?> as in this example:
1234 nigel 23 (?>\\d+)bar
1235 nigel 3
1236     This kind of parenthesis "locks up" the part of the pattern it contains once
1237     it has matched, and a failure further into the pattern is prevented from
1238     backtracking into it. Backtracking past it to previous items, however, works as
1239     normal.
1241 nigel 23 An alternative description is that a subpattern of this type matches the string
1242     of characters that an identical standalone pattern would match, if anchored at
1243     the current point in the subject string.
1244 nigel 3
1245 nigel 23 Once-only subpatterns are not capturing subpatterns. Simple cases such as the
1246 nigel 29 above example can be thought of as a maximizing repeat that must swallow
1247 nigel 23 everything it can. So, while both \\d+ and \\d+? are prepared to adjust the
1248     number of digits they match in order to make the rest of the pattern match,
1249     (?>\\d+) can only match an entire sequence of digits.
1251 nigel 3 This construction can of course contain arbitrarily complicated subpatterns,
1252 nigel 23 and it can be nested.
1253 nigel 3
1254 nigel 27 Once-only subpatterns can be used in conjunction with lookbehind assertions to
1255     specify efficient matching at the end of the subject string. Consider a simple
1256     pattern such as
1257 nigel 3
1258 nigel 27 abcd$
1260     when applied to a long string which does not match it. Because matching
1261     proceeds from left to right, PCRE will look for each "a" in the subject and
1262     then see if what follows matches the rest of the pattern. If the pattern is
1263     specified as
1265     .*abcd$
1267     then the initial .* matches the entire string at first, but when this fails, it
1268     backtracks to match all but the last character, then all but the last two
1269     characters, and so on. Once again the search for "a" covers the entire string,
1270     from right to left, so we are no better off. However, if the pattern is written
1271     as
1273     (?>.*)(?<=abcd)
1275     then there can be no backtracking for the .* item; it can match only the entire
1276     string. The subsequent lookbehind assertion does a single test on the last four
1277     characters. If it fails, the match fails immediately. For long strings, this
1278     approach makes a significant difference to the processing time.
1282     It is possible to cause the matching process to obey a subpattern
1283     conditionally or to choose between two alternative subpatterns, depending on
1284     the result of an assertion, or whether a previous capturing subpattern matched
1285     or not. The two possible forms of conditional subpattern are
1286 nigel 3
1287 nigel 23 (?(condition)yes-pattern)
1288     (?(condition)yes-pattern|no-pattern)
1289 nigel 3
1290 nigel 23 If the condition is satisfied, the yes-pattern is used; otherwise the
1291     no-pattern (if present) is used. If there are more than two alternatives in the
1292     subpattern, a compile-time error occurs.
1293 nigel 3
1294 nigel 23 There are two kinds of condition. If the text between the parentheses consists
1295     of a sequence of digits, then the condition is satisfied if the capturing
1296     subpattern of that number has previously matched. Consider the following
1297     pattern, which contains non-significant white space to make it more readable
1298     (assume the PCRE_EXTENDED option) and to divide it into three parts for ease
1299     of discussion:
1300 nigel 3
1301 nigel 23 ( \\( )? [^()]+ (?(1) \\) )
1302 nigel 3
1303 nigel 23 The first part matches an optional opening parenthesis, and if that
1304     character is present, sets it as the first captured substring. The second part
1305     matches one or more characters that are not parentheses. The third part is a
1306     conditional subpattern that tests whether the first set of parentheses matched
1307     or not. If they did, that is, if subject started with an opening parenthesis,
1308     the condition is true, and so the yes-pattern is executed and a closing
1309     parenthesis is required. Otherwise, since no-pattern is not present, the
1310     subpattern matches nothing. In other words, this pattern matches a sequence of
1311     non-parentheses, optionally enclosed in parentheses.
1312 nigel 3
1313 nigel 23 If the condition is not a sequence of digits, it must be an assertion. This may
1314     be a positive or negative lookahead or lookbehind assertion. Consider this
1315     pattern, again containing non-significant white space, and with the two
1316     alternatives on the second line:
1317 nigel 3
1318 nigel 23 (?(?=[^a-z]*[a-z])
1319     \\d{2}[a-z]{3}-\\d{2} | \\d{2}-\\d{2}-\\d{2} )
1320 nigel 19
1321 nigel 23 The condition is a positive lookahead assertion that matches an optional
1322     sequence of non-letters followed by a letter. In other words, it tests for the
1323     presence of at least one letter in the subject. If a letter is found, the
1324     subject is matched against the first alternative; otherwise it is matched
1325     against the second. This pattern matches strings in one of the two forms
1326     dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
1327 nigel 19
1328 nigel 3
1329 nigel 23 .SH COMMENTS
1330     The sequence (?# marks the start of a comment which continues up to the next
1331     closing parenthesis. Nested parentheses are not permitted. The characters
1332     that make up a comment play no part in the pattern matching at all.
1333 nigel 3
1334 nigel 23 If the PCRE_EXTENDED option is set, an unescaped # character outside a
1335     character class introduces a comment that continues up to the next newline
1336     character in the pattern.
1339 nigel 3 .SH PERFORMANCE
1340     Certain items that may appear in patterns are more efficient than others. It is
1341     more efficient to use a character class like [aeiou] than a set of alternatives
1342     such as (a|e|i|o|u). In general, the simplest construction that provides the
1343     required behaviour is usually the most efficient. Jeffrey Friedl's book
1344     contains a lot of discussion about optimizing regular expressions for efficient
1345     performance.
1348     .SH AUTHOR
1349     Philip Hazel <ph10@cam.ac.uk>
1350     .br
1351     University Computing Service,
1352     .br
1353     New Museums Site,
1354     .br
1355     Cambridge CB2 3QG, England.
1356     .br
1357     Phone: +44 1223 334714
1359 nigel 27 Copyright (c) 1997-1999 University of Cambridge.

ViewVC Help
Powered by ViewVC 1.1.12