1 | =head1 NAME |
---|
2 | |
---|
3 | perlre - Perl regular expressions |
---|
4 | |
---|
5 | =head1 DESCRIPTION |
---|
6 | |
---|
7 | This page describes the syntax of regular expressions in Perl. For a |
---|
8 | description of how to I<use> regular expressions in matching |
---|
9 | operations, plus various examples of the same, see discussions |
---|
10 | of C<m//>, C<s///>, C<qr//> and C<??> in L<perlop/"Regexp Quote-Like Operators">. |
---|
11 | |
---|
12 | Matching operations can have various modifiers. Modifiers |
---|
13 | that relate to the interpretation of the regular expression inside |
---|
14 | are listed below. Modifiers that alter the way a regular expression |
---|
15 | is used by Perl are detailed in L<perlop/"Regexp Quote-Like Operators"> and |
---|
16 | L<perlop/"Gory details of parsing quoted constructs">. |
---|
17 | |
---|
18 | =over 4 |
---|
19 | |
---|
20 | =item i |
---|
21 | |
---|
22 | Do case-insensitive pattern matching. |
---|
23 | |
---|
24 | If C<use locale> is in effect, the case map is taken from the current |
---|
25 | locale. See L<perllocale>. |
---|
26 | |
---|
27 | =item m |
---|
28 | |
---|
29 | Treat string as multiple lines. That is, change "^" and "$" from matching |
---|
30 | the start or end of the string to matching the start or end of any |
---|
31 | line anywhere within the string. |
---|
32 | |
---|
33 | =item s |
---|
34 | |
---|
35 | Treat string as single line. That is, change "." to match any character |
---|
36 | whatsoever, even a newline, which normally it would not match. |
---|
37 | |
---|
38 | The C</s> and C</m> modifiers both override the C<$*> setting. That |
---|
39 | is, no matter what C<$*> contains, C</s> without C</m> will force |
---|
40 | "^" to match only at the beginning of the string and "$" to match |
---|
41 | only at the end (or just before a newline at the end) of the string. |
---|
42 | Together, as /ms, they let the "." match any character whatsoever, |
---|
43 | while yet allowing "^" and "$" to match, respectively, just after |
---|
44 | and just before newlines within the string. |
---|
45 | |
---|
46 | =item x |
---|
47 | |
---|
48 | Extend your pattern's legibility by permitting whitespace and comments. |
---|
49 | |
---|
50 | =back |
---|
51 | |
---|
52 | These are usually written as "the C</x> modifier", even though the delimiter |
---|
53 | in question might not really be a slash. Any of these |
---|
54 | modifiers may also be embedded within the regular expression itself using |
---|
55 | the C<(?...)> construct. See below. |
---|
56 | |
---|
57 | The C</x> modifier itself needs a little more explanation. It tells |
---|
58 | the regular expression parser to ignore whitespace that is neither |
---|
59 | backslashed nor within a character class. You can use this to break up |
---|
60 | your regular expression into (slightly) more readable parts. The C<#> |
---|
61 | character is also treated as a metacharacter introducing a comment, |
---|
62 | just as in ordinary Perl code. This also means that if you want real |
---|
63 | whitespace or C<#> characters in the pattern (outside a character |
---|
64 | class, where they are unaffected by C</x>), that you'll either have to |
---|
65 | escape them or encode them using octal or hex escapes. Taken together, |
---|
66 | these features go a long way towards making Perl's regular expressions |
---|
67 | more readable. Note that you have to be careful not to include the |
---|
68 | pattern delimiter in the comment--perl has no way of knowing you did |
---|
69 | not intend to close the pattern early. See the C-comment deletion code |
---|
70 | in L<perlop>. |
---|
71 | |
---|
72 | =head2 Regular Expressions |
---|
73 | |
---|
74 | The patterns used in Perl pattern matching derive from supplied in |
---|
75 | the Version 8 regex routines. (The routines are derived |
---|
76 | (distantly) from Henry Spencer's freely redistributable reimplementation |
---|
77 | of the V8 routines.) See L<Version 8 Regular Expressions> for |
---|
78 | details. |
---|
79 | |
---|
80 | In particular the following metacharacters have their standard I<egrep>-ish |
---|
81 | meanings: |
---|
82 | |
---|
83 | \ Quote the next metacharacter |
---|
84 | ^ Match the beginning of the line |
---|
85 | . Match any character (except newline) |
---|
86 | $ Match the end of the line (or before newline at the end) |
---|
87 | | Alternation |
---|
88 | () Grouping |
---|
89 | [] Character class |
---|
90 | |
---|
91 | By default, the "^" character is guaranteed to match only the |
---|
92 | beginning of the string, the "$" character only the end (or before the |
---|
93 | newline at the end), and Perl does certain optimizations with the |
---|
94 | assumption that the string contains only one line. Embedded newlines |
---|
95 | will not be matched by "^" or "$". You may, however, wish to treat a |
---|
96 | string as a multi-line buffer, such that the "^" will match after any |
---|
97 | newline within the string, and "$" will match before any newline. At the |
---|
98 | cost of a little more overhead, you can do this by using the /m modifier |
---|
99 | on the pattern match operator. (Older programs did this by setting C<$*>, |
---|
100 | but this practice is now deprecated.) |
---|
101 | |
---|
102 | To simplify multi-line substitutions, the "." character never matches a |
---|
103 | newline unless you use the C</s> modifier, which in effect tells Perl to pretend |
---|
104 | the string is a single line--even if it isn't. The C</s> modifier also |
---|
105 | overrides the setting of C<$*>, in case you have some (badly behaved) older |
---|
106 | code that sets it in another module. |
---|
107 | |
---|
108 | The following standard quantifiers are recognized: |
---|
109 | |
---|
110 | * Match 0 or more times |
---|
111 | + Match 1 or more times |
---|
112 | ? Match 1 or 0 times |
---|
113 | {n} Match exactly n times |
---|
114 | {n,} Match at least n times |
---|
115 | {n,m} Match at least n but not more than m times |
---|
116 | |
---|
117 | (If a curly bracket occurs in any other context, it is treated |
---|
118 | as a regular character.) The "*" modifier is equivalent to C<{0,}>, the "+" |
---|
119 | modifier to C<{1,}>, and the "?" modifier to C<{0,1}>. n and m are limited |
---|
120 | to integral values less than a preset limit defined when perl is built. |
---|
121 | This is usually 32766 on the most common platforms. The actual limit can |
---|
122 | be seen in the error message generated by code such as this: |
---|
123 | |
---|
124 | $_ **= $_ , / {$_} / for 2 .. 42; |
---|
125 | |
---|
126 | By default, a quantified subpattern is "greedy", that is, it will match as |
---|
127 | many times as possible (given a particular starting location) while still |
---|
128 | allowing the rest of the pattern to match. If you want it to match the |
---|
129 | minimum number of times possible, follow the quantifier with a "?". Note |
---|
130 | that the meanings don't change, just the "greediness": |
---|
131 | |
---|
132 | *? Match 0 or more times |
---|
133 | +? Match 1 or more times |
---|
134 | ?? Match 0 or 1 time |
---|
135 | {n}? Match exactly n times |
---|
136 | {n,}? Match at least n times |
---|
137 | {n,m}? Match at least n but not more than m times |
---|
138 | |
---|
139 | Because patterns are processed as double quoted strings, the following |
---|
140 | also work: |
---|
141 | |
---|
142 | \t tab (HT, TAB) |
---|
143 | \n newline (LF, NL) |
---|
144 | \r return (CR) |
---|
145 | \f form feed (FF) |
---|
146 | \a alarm (bell) (BEL) |
---|
147 | \e escape (think troff) (ESC) |
---|
148 | \033 octal char (think of a PDP-11) |
---|
149 | \x1B hex char |
---|
150 | \x{263a} wide hex char (Unicode SMILEY) |
---|
151 | \c[ control char |
---|
152 | \N{name} named char |
---|
153 | \l lowercase next char (think vi) |
---|
154 | \u uppercase next char (think vi) |
---|
155 | \L lowercase till \E (think vi) |
---|
156 | \U uppercase till \E (think vi) |
---|
157 | \E end case modification (think vi) |
---|
158 | \Q quote (disable) pattern metacharacters till \E |
---|
159 | |
---|
160 | If C<use locale> is in effect, the case map used by C<\l>, C<\L>, C<\u> |
---|
161 | and C<\U> is taken from the current locale. See L<perllocale>. For |
---|
162 | documentation of C<\N{name}>, see L<charnames>. |
---|
163 | |
---|
164 | You cannot include a literal C<$> or C<@> within a C<\Q> sequence. |
---|
165 | An unescaped C<$> or C<@> interpolates the corresponding variable, |
---|
166 | while escaping will cause the literal string C<\$> to be matched. |
---|
167 | You'll need to write something like C<m/\Quser\E\@\Qhost/>. |
---|
168 | |
---|
169 | In addition, Perl defines the following: |
---|
170 | |
---|
171 | \w Match a "word" character (alphanumeric plus "_") |
---|
172 | \W Match a non-word character |
---|
173 | \s Match a whitespace character |
---|
174 | \S Match a non-whitespace character |
---|
175 | \d Match a digit character |
---|
176 | \D Match a non-digit character |
---|
177 | \pP Match P, named property. Use \p{Prop} for longer names. |
---|
178 | \PP Match non-P |
---|
179 | \X Match eXtended Unicode "combining character sequence", |
---|
180 | equivalent to C<(?:\PM\pM*)> |
---|
181 | \C Match a single C char (octet) even under utf8. |
---|
182 | |
---|
183 | A C<\w> matches a single alphanumeric character, not a whole word. |
---|
184 | Use C<\w+> to match a string of Perl-identifier characters (which isn't |
---|
185 | the same as matching an English word). If C<use locale> is in effect, the |
---|
186 | list of alphabetic characters generated by C<\w> is taken from the |
---|
187 | current locale. See L<perllocale>. You may use C<\w>, C<\W>, C<\s>, C<\S>, |
---|
188 | C<\d>, and C<\D> within character classes, but if you try to use them |
---|
189 | as endpoints of a range, that's not a range, the "-" is understood literally. |
---|
190 | See L<utf8> for details about C<\pP>, C<\PP>, and C<\X>. |
---|
191 | |
---|
192 | The POSIX character class syntax |
---|
193 | |
---|
194 | [:class:] |
---|
195 | |
---|
196 | is also available. The available classes and their backslash |
---|
197 | equivalents (if available) are as follows: |
---|
198 | |
---|
199 | alpha |
---|
200 | alnum |
---|
201 | ascii |
---|
202 | cntrl |
---|
203 | digit \d |
---|
204 | graph |
---|
205 | lower |
---|
206 | print |
---|
207 | punct |
---|
208 | space \s |
---|
209 | upper |
---|
210 | word \w |
---|
211 | xdigit |
---|
212 | |
---|
213 | For example use C<[:upper:]> to match all the uppercase characters. |
---|
214 | Note that the C<[]> are part of the C<[::]> construct, not part of the whole |
---|
215 | character class. For example: |
---|
216 | |
---|
217 | [01[:alpha:]%] |
---|
218 | |
---|
219 | matches one, zero, any alphabetic character, and the percentage sign. |
---|
220 | |
---|
221 | If the C<utf8> pragma is used, the following equivalences to Unicode |
---|
222 | \p{} constructs hold: |
---|
223 | |
---|
224 | alpha IsAlpha |
---|
225 | alnum IsAlnum |
---|
226 | ascii IsASCII |
---|
227 | cntrl IsCntrl |
---|
228 | digit IsDigit |
---|
229 | graph IsGraph |
---|
230 | lower IsLower |
---|
231 | print IsPrint |
---|
232 | punct IsPunct |
---|
233 | space IsSpace |
---|
234 | upper IsUpper |
---|
235 | word IsWord |
---|
236 | xdigit IsXDigit |
---|
237 | |
---|
238 | For example C<[:lower:]> and C<\p{IsLower}> are equivalent. |
---|
239 | |
---|
240 | If the C<utf8> pragma is not used but the C<locale> pragma is, the |
---|
241 | classes correlate with the isalpha(3) interface (except for `word', |
---|
242 | which is a Perl extension, mirroring C<\w>). |
---|
243 | |
---|
244 | The assumedly non-obviously named classes are: |
---|
245 | |
---|
246 | =over 4 |
---|
247 | |
---|
248 | =item cntrl |
---|
249 | |
---|
250 | Any control character. Usually characters that don't produce output as |
---|
251 | such but instead control the terminal somehow: for example newline and |
---|
252 | backspace are control characters. All characters with ord() less than |
---|
253 | 32 are most often classified as control characters. |
---|
254 | |
---|
255 | =item graph |
---|
256 | |
---|
257 | Any alphanumeric or punctuation character. |
---|
258 | |
---|
259 | =item print |
---|
260 | |
---|
261 | Any alphanumeric or punctuation character or space. |
---|
262 | |
---|
263 | =item punct |
---|
264 | |
---|
265 | Any punctuation character. |
---|
266 | |
---|
267 | =item xdigit |
---|
268 | |
---|
269 | Any hexadecimal digit. Though this may feel silly (/0-9a-f/i would |
---|
270 | work just fine) it is included for completeness. |
---|
271 | |
---|
272 | =back |
---|
273 | |
---|
274 | You can negate the [::] character classes by prefixing the class name |
---|
275 | with a '^'. This is a Perl extension. For example: |
---|
276 | |
---|
277 | POSIX trad. Perl utf8 Perl |
---|
278 | |
---|
279 | [:^digit:] \D \P{IsDigit} |
---|
280 | [:^space:] \S \P{IsSpace} |
---|
281 | [:^word:] \W \P{IsWord} |
---|
282 | |
---|
283 | The POSIX character classes [.cc.] and [=cc=] are recognized but |
---|
284 | B<not> supported and trying to use them will cause an error. |
---|
285 | |
---|
286 | Perl defines the following zero-width assertions: |
---|
287 | |
---|
288 | \b Match a word boundary |
---|
289 | \B Match a non-(word boundary) |
---|
290 | \A Match only at beginning of string |
---|
291 | \Z Match only at end of string, or before newline at the end |
---|
292 | \z Match only at end of string |
---|
293 | \G Match only at pos() (e.g. at the end-of-match position |
---|
294 | of prior m//g) |
---|
295 | |
---|
296 | A word boundary (C<\b>) is a spot between two characters |
---|
297 | that has a C<\w> on one side of it and a C<\W> on the other side |
---|
298 | of it (in either order), counting the imaginary characters off the |
---|
299 | beginning and end of the string as matching a C<\W>. (Within |
---|
300 | character classes C<\b> represents backspace rather than a word |
---|
301 | boundary, just as it normally does in any double-quoted string.) |
---|
302 | The C<\A> and C<\Z> are just like "^" and "$", except that they |
---|
303 | won't match multiple times when the C</m> modifier is used, while |
---|
304 | "^" and "$" will match at every internal line boundary. To match |
---|
305 | the actual end of the string and not ignore an optional trailing |
---|
306 | newline, use C<\z>. |
---|
307 | |
---|
308 | The C<\G> assertion can be used to chain global matches (using |
---|
309 | C<m//g>), as described in L<perlop/"Regexp Quote-Like Operators">. |
---|
310 | It is also useful when writing C<lex>-like scanners, when you have |
---|
311 | several patterns that you want to match against consequent substrings |
---|
312 | of your string, see the previous reference. The actual location |
---|
313 | where C<\G> will match can also be influenced by using C<pos()> as |
---|
314 | an lvalue. See L<perlfunc/pos>. |
---|
315 | |
---|
316 | The bracketing construct C<( ... )> creates capture buffers. To |
---|
317 | refer to the digit'th buffer use \<digit> within the |
---|
318 | match. Outside the match use "$" instead of "\". (The |
---|
319 | \<digit> notation works in certain circumstances outside |
---|
320 | the match. See the warning below about \1 vs $1 for details.) |
---|
321 | Referring back to another part of the match is called a |
---|
322 | I<backreference>. |
---|
323 | |
---|
324 | There is no limit to the number of captured substrings that you may |
---|
325 | use. However Perl also uses \10, \11, etc. as aliases for \010, |
---|
326 | \011, etc. (Recall that 0 means octal, so \011 is the 9'th ASCII |
---|
327 | character, a tab.) Perl resolves this ambiguity by interpreting |
---|
328 | \10 as a backreference only if at least 10 left parentheses have |
---|
329 | opened before it. Likewise \11 is a backreference only if at least |
---|
330 | 11 left parentheses have opened before it. And so on. \1 through |
---|
331 | \9 are always interpreted as backreferences." |
---|
332 | |
---|
333 | Examples: |
---|
334 | |
---|
335 | s/^([^ ]*) *([^ ]*)/$2 $1/; # swap first two words |
---|
336 | |
---|
337 | if (/(.)\1/) { # find first doubled char |
---|
338 | print "'$1' is the first doubled character\n"; |
---|
339 | } |
---|
340 | |
---|
341 | if (/Time: (..):(..):(..)/) { # parse out values |
---|
342 | $hours = $1; |
---|
343 | $minutes = $2; |
---|
344 | $seconds = $3; |
---|
345 | } |
---|
346 | |
---|
347 | Several special variables also refer back to portions of the previous |
---|
348 | match. C<$+> returns whatever the last bracket match matched. |
---|
349 | C<$&> returns the entire matched string. (At one point C<$0> did |
---|
350 | also, but now it returns the name of the program.) C<$`> returns |
---|
351 | everything before the matched string. And C<$'> returns everything |
---|
352 | after the matched string. |
---|
353 | |
---|
354 | The numbered variables ($1, $2, $3, etc.) and the related punctuation |
---|
355 | set (C<<$+>, C<$&>, C<$`>, and C<$'>) are all dynamically scoped |
---|
356 | until the end of the enclosing block or until the next successful |
---|
357 | match, whichever comes first. (See L<perlsyn/"Compound Statements">.) |
---|
358 | |
---|
359 | B<WARNING>: Once Perl sees that you need one of C<$&>, C<$`>, or |
---|
360 | C<$'> anywhere in the program, it has to provide them for every |
---|
361 | pattern match. This may substantially slow your program. Perl |
---|
362 | uses the same mechanism to produce $1, $2, etc, so you also pay a |
---|
363 | price for each pattern that contains capturing parentheses. (To |
---|
364 | avoid this cost while retaining the grouping behaviour, use the |
---|
365 | extended regular expression C<(?: ... )> instead.) But if you never |
---|
366 | use C<$&>, C<$`> or C<$'>, then patterns I<without> capturing |
---|
367 | parentheses will not be penalized. So avoid C<$&>, C<$'>, and C<$`> |
---|
368 | if you can, but if you can't (and some algorithms really appreciate |
---|
369 | them), once you've used them once, use them at will, because you've |
---|
370 | already paid the price. As of 5.005, C<$&> is not so costly as the |
---|
371 | other two. |
---|
372 | |
---|
373 | Backslashed metacharacters in Perl are alphanumeric, such as C<\b>, |
---|
374 | C<\w>, C<\n>. Unlike some other regular expression languages, there |
---|
375 | are no backslashed symbols that aren't alphanumeric. So anything |
---|
376 | that looks like \\, \(, \), \<, \>, \{, or \} is always |
---|
377 | interpreted as a literal character, not a metacharacter. This was |
---|
378 | once used in a common idiom to disable or quote the special meanings |
---|
379 | of regular expression metacharacters in a string that you want to |
---|
380 | use for a pattern. Simply quote all non-alphanumeric characters: |
---|
381 | |
---|
382 | $pattern =~ s/(\W)/\\$1/g; |
---|
383 | |
---|
384 | Today it is more common to use the quotemeta() function or the C<\Q> |
---|
385 | metaquoting escape sequence to disable all metacharacters' special |
---|
386 | meanings like this: |
---|
387 | |
---|
388 | /$unquoted\Q$quoted\E$unquoted/ |
---|
389 | |
---|
390 | Beware that if you put literal backslashes (those not inside |
---|
391 | interpolated variables) between C<\Q> and C<\E>, double-quotish |
---|
392 | backslash interpolation may lead to confusing results. If you |
---|
393 | I<need> to use literal backslashes within C<\Q...\E>, |
---|
394 | consult L<perlop/"Gory details of parsing quoted constructs">. |
---|
395 | |
---|
396 | =head2 Extended Patterns |
---|
397 | |
---|
398 | Perl also defines a consistent extension syntax for features not |
---|
399 | found in standard tools like B<awk> and B<lex>. The syntax is a |
---|
400 | pair of parentheses with a question mark as the first thing within |
---|
401 | the parentheses. The character after the question mark indicates |
---|
402 | the extension. |
---|
403 | |
---|
404 | The stability of these extensions varies widely. Some have been |
---|
405 | part of the core language for many years. Others are experimental |
---|
406 | and may change without warning or be completely removed. Check |
---|
407 | the documentation on an individual feature to verify its current |
---|
408 | status. |
---|
409 | |
---|
410 | A question mark was chosen for this and for the minimal-matching |
---|
411 | construct because 1) question marks are rare in older regular |
---|
412 | expressions, and 2) whenever you see one, you should stop and |
---|
413 | "question" exactly what is going on. That's psychology... |
---|
414 | |
---|
415 | =over 10 |
---|
416 | |
---|
417 | =item C<(?#text)> |
---|
418 | |
---|
419 | A comment. The text is ignored. If the C</x> modifier enables |
---|
420 | whitespace formatting, a simple C<#> will suffice. Note that Perl closes |
---|
421 | the comment as soon as it sees a C<)>, so there is no way to put a literal |
---|
422 | C<)> in the comment. |
---|
423 | |
---|
424 | =item C<(?imsx-imsx)> |
---|
425 | |
---|
426 | One or more embedded pattern-match modifiers. This is particularly |
---|
427 | useful for dynamic patterns, such as those read in from a configuration |
---|
428 | file, read in as an argument, are specified in a table somewhere, |
---|
429 | etc. Consider the case that some of which want to be case sensitive |
---|
430 | and some do not. The case insensitive ones need to include merely |
---|
431 | C<(?i)> at the front of the pattern. For example: |
---|
432 | |
---|
433 | $pattern = "foobar"; |
---|
434 | if ( /$pattern/i ) { } |
---|
435 | |
---|
436 | # more flexible: |
---|
437 | |
---|
438 | $pattern = "(?i)foobar"; |
---|
439 | if ( /$pattern/ ) { } |
---|
440 | |
---|
441 | Letters after a C<-> turn those modifiers off. These modifiers are |
---|
442 | localized inside an enclosing group (if any). For example, |
---|
443 | |
---|
444 | ( (?i) blah ) \s+ \1 |
---|
445 | |
---|
446 | will match a repeated (I<including the case>!) word C<blah> in any |
---|
447 | case, assuming C<x> modifier, and no C<i> modifier outside this |
---|
448 | group. |
---|
449 | |
---|
450 | =item C<(?:pattern)> |
---|
451 | |
---|
452 | =item C<(?imsx-imsx:pattern)> |
---|
453 | |
---|
454 | This is for clustering, not capturing; it groups subexpressions like |
---|
455 | "()", but doesn't make backreferences as "()" does. So |
---|
456 | |
---|
457 | @fields = split(/\b(?:a|b|c)\b/) |
---|
458 | |
---|
459 | is like |
---|
460 | |
---|
461 | @fields = split(/\b(a|b|c)\b/) |
---|
462 | |
---|
463 | but doesn't spit out extra fields. It's also cheaper not to capture |
---|
464 | characters if you don't need to. |
---|
465 | |
---|
466 | Any letters between C<?> and C<:> act as flags modifiers as with |
---|
467 | C<(?imsx-imsx)>. For example, |
---|
468 | |
---|
469 | /(?s-i:more.*than).*million/i |
---|
470 | |
---|
471 | is equivalent to the more verbose |
---|
472 | |
---|
473 | /(?:(?s-i)more.*than).*million/i |
---|
474 | |
---|
475 | =item C<(?=pattern)> |
---|
476 | |
---|
477 | A zero-width positive look-ahead assertion. For example, C</\w+(?=\t)/> |
---|
478 | matches a word followed by a tab, without including the tab in C<$&>. |
---|
479 | |
---|
480 | =item C<(?!pattern)> |
---|
481 | |
---|
482 | A zero-width negative look-ahead assertion. For example C</foo(?!bar)/> |
---|
483 | matches any occurrence of "foo" that isn't followed by "bar". Note |
---|
484 | however that look-ahead and look-behind are NOT the same thing. You cannot |
---|
485 | use this for look-behind. |
---|
486 | |
---|
487 | If you are looking for a "bar" that isn't preceded by a "foo", C</(?!foo)bar/> |
---|
488 | will not do what you want. That's because the C<(?!foo)> is just saying that |
---|
489 | the next thing cannot be "foo"--and it's not, it's a "bar", so "foobar" will |
---|
490 | match. You would have to do something like C</(?!foo)...bar/> for that. We |
---|
491 | say "like" because there's the case of your "bar" not having three characters |
---|
492 | before it. You could cover that this way: C</(?:(?!foo)...|^.{0,2})bar/>. |
---|
493 | Sometimes it's still easier just to say: |
---|
494 | |
---|
495 | if (/bar/ && $` !~ /foo$/) |
---|
496 | |
---|
497 | For look-behind see below. |
---|
498 | |
---|
499 | =item C<(?<=pattern)> |
---|
500 | |
---|
501 | A zero-width positive look-behind assertion. For example, C</(?<=\t)\w+/> |
---|
502 | matches a word that follows a tab, without including the tab in C<$&>. |
---|
503 | Works only for fixed-width look-behind. |
---|
504 | |
---|
505 | =item C<(?<!pattern)> |
---|
506 | |
---|
507 | A zero-width negative look-behind assertion. For example C</(?<!bar)foo/> |
---|
508 | matches any occurrence of "foo" that does not follow "bar". Works |
---|
509 | only for fixed-width look-behind. |
---|
510 | |
---|
511 | =item C<(?{ code })> |
---|
512 | |
---|
513 | B<WARNING>: This extended regular expression feature is considered |
---|
514 | highly experimental, and may be changed or deleted without notice. |
---|
515 | |
---|
516 | This zero-width assertion evaluate any embedded Perl code. It |
---|
517 | always succeeds, and its C<code> is not interpolated. Currently, |
---|
518 | the rules to determine where the C<code> ends are somewhat convoluted. |
---|
519 | |
---|
520 | The C<code> is properly scoped in the following sense: If the assertion |
---|
521 | is backtracked (compare L<"Backtracking">), all changes introduced after |
---|
522 | C<local>ization are undone, so that |
---|
523 | |
---|
524 | $_ = 'a' x 8; |
---|
525 | m< |
---|
526 | (?{ $cnt = 0 }) # Initialize $cnt. |
---|
527 | ( |
---|
528 | a |
---|
529 | (?{ |
---|
530 | local $cnt = $cnt + 1; # Update $cnt, backtracking-safe. |
---|
531 | }) |
---|
532 | )* |
---|
533 | aaaa |
---|
534 | (?{ $res = $cnt }) # On success copy to non-localized |
---|
535 | # location. |
---|
536 | >x; |
---|
537 | |
---|
538 | will set C<$res = 4>. Note that after the match, $cnt returns to the globally |
---|
539 | introduced value, because the scopes that restrict C<local> operators |
---|
540 | are unwound. |
---|
541 | |
---|
542 | This assertion may be used as a C<(?(condition)yes-pattern|no-pattern)> |
---|
543 | switch. If I<not> used in this way, the result of evaluation of |
---|
544 | C<code> is put into the special variable C<$^R>. This happens |
---|
545 | immediately, so C<$^R> can be used from other C<(?{ code })> assertions |
---|
546 | inside the same regular expression. |
---|
547 | |
---|
548 | The assignment to C<$^R> above is properly localized, so the old |
---|
549 | value of C<$^R> is restored if the assertion is backtracked; compare |
---|
550 | L<"Backtracking">. |
---|
551 | |
---|
552 | For reasons of security, this construct is forbidden if the regular |
---|
553 | expression involves run-time interpolation of variables, unless the |
---|
554 | perilous C<use re 'eval'> pragma has been used (see L<re>), or the |
---|
555 | variables contain results of C<qr//> operator (see |
---|
556 | L<perlop/"qr/STRING/imosx">). |
---|
557 | |
---|
558 | This restriction is because of the wide-spread and remarkably convenient |
---|
559 | custom of using run-time determined strings as patterns. For example: |
---|
560 | |
---|
561 | $re = <>; |
---|
562 | chomp $re; |
---|
563 | $string =~ /$re/; |
---|
564 | |
---|
565 | Before Perl knew how to execute interpolated code within a pattern, |
---|
566 | this operation was completely safe from a security point of view, |
---|
567 | although it could raise an exception from an illegal pattern. If |
---|
568 | you turn on the C<use re 'eval'>, though, it is no longer secure, |
---|
569 | so you should only do so if you are also using taint checking. |
---|
570 | Better yet, use the carefully constrained evaluation within a Safe |
---|
571 | module. See L<perlsec> for details about both these mechanisms. |
---|
572 | |
---|
573 | =item C<(??{ code })> |
---|
574 | |
---|
575 | B<WARNING>: This extended regular expression feature is considered |
---|
576 | highly experimental, and may be changed or deleted without notice. |
---|
577 | A simplified version of the syntax may be introduced for commonly |
---|
578 | used idioms. |
---|
579 | |
---|
580 | This is a "postponed" regular subexpression. The C<code> is evaluated |
---|
581 | at run time, at the moment this subexpression may match. The result |
---|
582 | of evaluation is considered as a regular expression and matched as |
---|
583 | if it were inserted instead of this construct. |
---|
584 | |
---|
585 | The C<code> is not interpolated. As before, the rules to determine |
---|
586 | where the C<code> ends are currently somewhat convoluted. |
---|
587 | |
---|
588 | The following pattern matches a parenthesized group: |
---|
589 | |
---|
590 | $re = qr{ |
---|
591 | \( |
---|
592 | (?: |
---|
593 | (?> [^()]+ ) # Non-parens without backtracking |
---|
594 | | |
---|
595 | (??{ $re }) # Group with matching parens |
---|
596 | )* |
---|
597 | \) |
---|
598 | }x; |
---|
599 | |
---|
600 | =item C<< (?>pattern) >> |
---|
601 | |
---|
602 | B<WARNING>: This extended regular expression feature is considered |
---|
603 | highly experimental, and may be changed or deleted without notice. |
---|
604 | |
---|
605 | An "independent" subexpression, one which matches the substring |
---|
606 | that a I<standalone> C<pattern> would match if anchored at the given |
---|
607 | position, and it matches I<nothing other than this substring>. This |
---|
608 | construct is useful for optimizations of what would otherwise be |
---|
609 | "eternal" matches, because it will not backtrack (see L<"Backtracking">). |
---|
610 | It may also be useful in places where the "grab all you can, and do not |
---|
611 | give anything back" semantic is desirable. |
---|
612 | |
---|
613 | For example: C<< ^(?>a*)ab >> will never match, since C<< (?>a*) >> |
---|
614 | (anchored at the beginning of string, as above) will match I<all> |
---|
615 | characters C<a> at the beginning of string, leaving no C<a> for |
---|
616 | C<ab> to match. In contrast, C<a*ab> will match the same as C<a+b>, |
---|
617 | since the match of the subgroup C<a*> is influenced by the following |
---|
618 | group C<ab> (see L<"Backtracking">). In particular, C<a*> inside |
---|
619 | C<a*ab> will match fewer characters than a standalone C<a*>, since |
---|
620 | this makes the tail match. |
---|
621 | |
---|
622 | An effect similar to C<< (?>pattern) >> may be achieved by writing |
---|
623 | C<(?=(pattern))\1>. This matches the same substring as a standalone |
---|
624 | C<a+>, and the following C<\1> eats the matched string; it therefore |
---|
625 | makes a zero-length assertion into an analogue of C<< (?>...) >>. |
---|
626 | (The difference between these two constructs is that the second one |
---|
627 | uses a capturing group, thus shifting ordinals of backreferences |
---|
628 | in the rest of a regular expression.) |
---|
629 | |
---|
630 | Consider this pattern: |
---|
631 | |
---|
632 | m{ \( |
---|
633 | ( |
---|
634 | [^()]+ # x+ |
---|
635 | | |
---|
636 | \( [^()]* \) |
---|
637 | )+ |
---|
638 | \) |
---|
639 | }x |
---|
640 | |
---|
641 | That will efficiently match a nonempty group with matching parentheses |
---|
642 | two levels deep or less. However, if there is no such group, it |
---|
643 | will take virtually forever on a long string. That's because there |
---|
644 | are so many different ways to split a long string into several |
---|
645 | substrings. This is what C<(.+)+> is doing, and C<(.+)+> is similar |
---|
646 | to a subpattern of the above pattern. Consider how the pattern |
---|
647 | above detects no-match on C<((()aaaaaaaaaaaaaaaaaa> in several |
---|
648 | seconds, but that each extra letter doubles this time. This |
---|
649 | exponential performance will make it appear that your program has |
---|
650 | hung. However, a tiny change to this pattern |
---|
651 | |
---|
652 | m{ \( |
---|
653 | ( |
---|
654 | (?> [^()]+ ) # change x+ above to (?> x+ ) |
---|
655 | | |
---|
656 | \( [^()]* \) |
---|
657 | )+ |
---|
658 | \) |
---|
659 | }x |
---|
660 | |
---|
661 | which uses C<< (?>...) >> matches exactly when the one above does (verifying |
---|
662 | this yourself would be a productive exercise), but finishes in a fourth |
---|
663 | the time when used on a similar string with 1000000 C<a>s. Be aware, |
---|
664 | however, that this pattern currently triggers a warning message under |
---|
665 | the C<use warnings> pragma or B<-w> switch saying it |
---|
666 | C<"matches the null string many times">): |
---|
667 | |
---|
668 | On simple groups, such as the pattern C<< (?> [^()]+ ) >>, a comparable |
---|
669 | effect may be achieved by negative look-ahead, as in C<[^()]+ (?! [^()] )>. |
---|
670 | This was only 4 times slower on a string with 1000000 C<a>s. |
---|
671 | |
---|
672 | The "grab all you can, and do not give anything back" semantic is desirable |
---|
673 | in many situations where on the first sight a simple C<()*> looks like |
---|
674 | the correct solution. Suppose we parse text with comments being delimited |
---|
675 | by C<#> followed by some optional (horizontal) whitespace. Contrary to |
---|
676 | its appearence, C<#[ \t]*> I<is not> the correct subexpression to match |
---|
677 | the comment delimiter, because it may "give up" some whitespace if |
---|
678 | the remainder of the pattern can be made to match that way. The correct |
---|
679 | answer is either one of these: |
---|
680 | |
---|
681 | (?>#[ \t]*) |
---|
682 | #[ \t]*(?![ \t]) |
---|
683 | |
---|
684 | For example, to grab non-empty comments into $1, one should use either |
---|
685 | one of these: |
---|
686 | |
---|
687 | / (?> \# [ \t]* ) ( .+ ) /x; |
---|
688 | / \# [ \t]* ( [^ \t] .* ) /x; |
---|
689 | |
---|
690 | Which one you pick depends on which of these expressions better reflects |
---|
691 | the above specification of comments. |
---|
692 | |
---|
693 | =item C<(?(condition)yes-pattern|no-pattern)> |
---|
694 | |
---|
695 | =item C<(?(condition)yes-pattern)> |
---|
696 | |
---|
697 | B<WARNING>: This extended regular expression feature is considered |
---|
698 | highly experimental, and may be changed or deleted without notice. |
---|
699 | |
---|
700 | Conditional expression. C<(condition)> should be either an integer in |
---|
701 | parentheses (which is valid if the corresponding pair of parentheses |
---|
702 | matched), or look-ahead/look-behind/evaluate zero-width assertion. |
---|
703 | |
---|
704 | For example: |
---|
705 | |
---|
706 | m{ ( \( )? |
---|
707 | [^()]+ |
---|
708 | (?(1) \) ) |
---|
709 | }x |
---|
710 | |
---|
711 | matches a chunk of non-parentheses, possibly included in parentheses |
---|
712 | themselves. |
---|
713 | |
---|
714 | =back |
---|
715 | |
---|
716 | =head2 Backtracking |
---|
717 | |
---|
718 | NOTE: This section presents an abstract approximation of regular |
---|
719 | expression behavior. For a more rigorous (and complicated) view of |
---|
720 | the rules involved in selecting a match among possible alternatives, |
---|
721 | see L<Combining pieces together>. |
---|
722 | |
---|
723 | A fundamental feature of regular expression matching involves the |
---|
724 | notion called I<backtracking>, which is currently used (when needed) |
---|
725 | by all regular expression quantifiers, namely C<*>, C<*?>, C<+>, |
---|
726 | C<+?>, C<{n,m}>, and C<{n,m}?>. Backtracking is often optimized |
---|
727 | internally, but the general principle outlined here is valid. |
---|
728 | |
---|
729 | For a regular expression to match, the I<entire> regular expression must |
---|
730 | match, not just part of it. So if the beginning of a pattern containing a |
---|
731 | quantifier succeeds in a way that causes later parts in the pattern to |
---|
732 | fail, the matching engine backs up and recalculates the beginning |
---|
733 | part--that's why it's called backtracking. |
---|
734 | |
---|
735 | Here is an example of backtracking: Let's say you want to find the |
---|
736 | word following "foo" in the string "Food is on the foo table.": |
---|
737 | |
---|
738 | $_ = "Food is on the foo table."; |
---|
739 | if ( /\b(foo)\s+(\w+)/i ) { |
---|
740 | print "$2 follows $1.\n"; |
---|
741 | } |
---|
742 | |
---|
743 | When the match runs, the first part of the regular expression (C<\b(foo)>) |
---|
744 | finds a possible match right at the beginning of the string, and loads up |
---|
745 | $1 with "Foo". However, as soon as the matching engine sees that there's |
---|
746 | no whitespace following the "Foo" that it had saved in $1, it realizes its |
---|
747 | mistake and starts over again one character after where it had the |
---|
748 | tentative match. This time it goes all the way until the next occurrence |
---|
749 | of "foo". The complete regular expression matches this time, and you get |
---|
750 | the expected output of "table follows foo." |
---|
751 | |
---|
752 | Sometimes minimal matching can help a lot. Imagine you'd like to match |
---|
753 | everything between "foo" and "bar". Initially, you write something |
---|
754 | like this: |
---|
755 | |
---|
756 | $_ = "The food is under the bar in the barn."; |
---|
757 | if ( /foo(.*)bar/ ) { |
---|
758 | print "got <$1>\n"; |
---|
759 | } |
---|
760 | |
---|
761 | Which perhaps unexpectedly yields: |
---|
762 | |
---|
763 | got <d is under the bar in the > |
---|
764 | |
---|
765 | That's because C<.*> was greedy, so you get everything between the |
---|
766 | I<first> "foo" and the I<last> "bar". Here it's more effective |
---|
767 | to use minimal matching to make sure you get the text between a "foo" |
---|
768 | and the first "bar" thereafter. |
---|
769 | |
---|
770 | if ( /foo(.*?)bar/ ) { print "got <$1>\n" } |
---|
771 | got <d is under the > |
---|
772 | |
---|
773 | Here's another example: let's say you'd like to match a number at the end |
---|
774 | of a string, and you also want to keep the preceding part the match. |
---|
775 | So you write this: |
---|
776 | |
---|
777 | $_ = "I have 2 numbers: 53147"; |
---|
778 | if ( /(.*)(\d*)/ ) { # Wrong! |
---|
779 | print "Beginning is <$1>, number is <$2>.\n"; |
---|
780 | } |
---|
781 | |
---|
782 | That won't work at all, because C<.*> was greedy and gobbled up the |
---|
783 | whole string. As C<\d*> can match on an empty string the complete |
---|
784 | regular expression matched successfully. |
---|
785 | |
---|
786 | Beginning is <I have 2 numbers: 53147>, number is <>. |
---|
787 | |
---|
788 | Here are some variants, most of which don't work: |
---|
789 | |
---|
790 | $_ = "I have 2 numbers: 53147"; |
---|
791 | @pats = qw{ |
---|
792 | (.*)(\d*) |
---|
793 | (.*)(\d+) |
---|
794 | (.*?)(\d*) |
---|
795 | (.*?)(\d+) |
---|
796 | (.*)(\d+)$ |
---|
797 | (.*?)(\d+)$ |
---|
798 | (.*)\b(\d+)$ |
---|
799 | (.*\D)(\d+)$ |
---|
800 | }; |
---|
801 | |
---|
802 | for $pat (@pats) { |
---|
803 | printf "%-12s ", $pat; |
---|
804 | if ( /$pat/ ) { |
---|
805 | print "<$1> <$2>\n"; |
---|
806 | } else { |
---|
807 | print "FAIL\n"; |
---|
808 | } |
---|
809 | } |
---|
810 | |
---|
811 | That will print out: |
---|
812 | |
---|
813 | (.*)(\d*) <I have 2 numbers: 53147> <> |
---|
814 | (.*)(\d+) <I have 2 numbers: 5314> <7> |
---|
815 | (.*?)(\d*) <> <> |
---|
816 | (.*?)(\d+) <I have > <2> |
---|
817 | (.*)(\d+)$ <I have 2 numbers: 5314> <7> |
---|
818 | (.*?)(\d+)$ <I have 2 numbers: > <53147> |
---|
819 | (.*)\b(\d+)$ <I have 2 numbers: > <53147> |
---|
820 | (.*\D)(\d+)$ <I have 2 numbers: > <53147> |
---|
821 | |
---|
822 | As you see, this can be a bit tricky. It's important to realize that a |
---|
823 | regular expression is merely a set of assertions that gives a definition |
---|
824 | of success. There may be 0, 1, or several different ways that the |
---|
825 | definition might succeed against a particular string. And if there are |
---|
826 | multiple ways it might succeed, you need to understand backtracking to |
---|
827 | know which variety of success you will achieve. |
---|
828 | |
---|
829 | When using look-ahead assertions and negations, this can all get even |
---|
830 | tricker. Imagine you'd like to find a sequence of non-digits not |
---|
831 | followed by "123". You might try to write that as |
---|
832 | |
---|
833 | $_ = "ABC123"; |
---|
834 | if ( /^\D*(?!123)/ ) { # Wrong! |
---|
835 | print "Yup, no 123 in $_\n"; |
---|
836 | } |
---|
837 | |
---|
838 | But that isn't going to match; at least, not the way you're hoping. It |
---|
839 | claims that there is no 123 in the string. Here's a clearer picture of |
---|
840 | why it that pattern matches, contrary to popular expectations: |
---|
841 | |
---|
842 | $x = 'ABC123' ; |
---|
843 | $y = 'ABC445' ; |
---|
844 | |
---|
845 | print "1: got $1\n" if $x =~ /^(ABC)(?!123)/ ; |
---|
846 | print "2: got $1\n" if $y =~ /^(ABC)(?!123)/ ; |
---|
847 | |
---|
848 | print "3: got $1\n" if $x =~ /^(\D*)(?!123)/ ; |
---|
849 | print "4: got $1\n" if $y =~ /^(\D*)(?!123)/ ; |
---|
850 | |
---|
851 | This prints |
---|
852 | |
---|
853 | 2: got ABC |
---|
854 | 3: got AB |
---|
855 | 4: got ABC |
---|
856 | |
---|
857 | You might have expected test 3 to fail because it seems to a more |
---|
858 | general purpose version of test 1. The important difference between |
---|
859 | them is that test 3 contains a quantifier (C<\D*>) and so can use |
---|
860 | backtracking, whereas test 1 will not. What's happening is |
---|
861 | that you've asked "Is it true that at the start of $x, following 0 or more |
---|
862 | non-digits, you have something that's not 123?" If the pattern matcher had |
---|
863 | let C<\D*> expand to "ABC", this would have caused the whole pattern to |
---|
864 | fail. |
---|
865 | |
---|
866 | The search engine will initially match C<\D*> with "ABC". Then it will |
---|
867 | try to match C<(?!123> with "123", which fails. But because |
---|
868 | a quantifier (C<\D*>) has been used in the regular expression, the |
---|
869 | search engine can backtrack and retry the match differently |
---|
870 | in the hope of matching the complete regular expression. |
---|
871 | |
---|
872 | The pattern really, I<really> wants to succeed, so it uses the |
---|
873 | standard pattern back-off-and-retry and lets C<\D*> expand to just "AB" this |
---|
874 | time. Now there's indeed something following "AB" that is not |
---|
875 | "123". It's "C123", which suffices. |
---|
876 | |
---|
877 | We can deal with this by using both an assertion and a negation. |
---|
878 | We'll say that the first part in $1 must be followed both by a digit |
---|
879 | and by something that's not "123". Remember that the look-aheads |
---|
880 | are zero-width expressions--they only look, but don't consume any |
---|
881 | of the string in their match. So rewriting this way produces what |
---|
882 | you'd expect; that is, case 5 will fail, but case 6 succeeds: |
---|
883 | |
---|
884 | print "5: got $1\n" if $x =~ /^(\D*)(?=\d)(?!123)/ ; |
---|
885 | print "6: got $1\n" if $y =~ /^(\D*)(?=\d)(?!123)/ ; |
---|
886 | |
---|
887 | 6: got ABC |
---|
888 | |
---|
889 | In other words, the two zero-width assertions next to each other work as though |
---|
890 | they're ANDed together, just as you'd use any built-in assertions: C</^$/> |
---|
891 | matches only if you're at the beginning of the line AND the end of the |
---|
892 | line simultaneously. The deeper underlying truth is that juxtaposition in |
---|
893 | regular expressions always means AND, except when you write an explicit OR |
---|
894 | using the vertical bar. C</ab/> means match "a" AND (then) match "b", |
---|
895 | although the attempted matches are made at different positions because "a" |
---|
896 | is not a zero-width assertion, but a one-width assertion. |
---|
897 | |
---|
898 | B<WARNING>: particularly complicated regular expressions can take |
---|
899 | exponential time to solve because of the immense number of possible |
---|
900 | ways they can use backtracking to try match. For example, without |
---|
901 | internal optimizations done by the regular expression engine, this will |
---|
902 | take a painfully long time to run: |
---|
903 | |
---|
904 | 'aaaaaaaaaaaa' =~ /((a{0,5}){0,5}){0,5}[c]/ |
---|
905 | |
---|
906 | And if you used C<*>'s instead of limiting it to 0 through 5 matches, |
---|
907 | then it would take forever--or until you ran out of stack space. |
---|
908 | |
---|
909 | A powerful tool for optimizing such beasts is what is known as an |
---|
910 | "independent group", |
---|
911 | which does not backtrack (see L<C<< (?>pattern) >>>). Note also that |
---|
912 | zero-length look-ahead/look-behind assertions will not backtrack to make |
---|
913 | the tail match, since they are in "logical" context: only |
---|
914 | whether they match is considered relevant. For an example |
---|
915 | where side-effects of look-ahead I<might> have influenced the |
---|
916 | following match, see L<C<< (?>pattern) >>>. |
---|
917 | |
---|
918 | =head2 Version 8 Regular Expressions |
---|
919 | |
---|
920 | In case you're not familiar with the "regular" Version 8 regex |
---|
921 | routines, here are the pattern-matching rules not described above. |
---|
922 | |
---|
923 | Any single character matches itself, unless it is a I<metacharacter> |
---|
924 | with a special meaning described here or above. You can cause |
---|
925 | characters that normally function as metacharacters to be interpreted |
---|
926 | literally by prefixing them with a "\" (e.g., "\." matches a ".", not any |
---|
927 | character; "\\" matches a "\"). A series of characters matches that |
---|
928 | series of characters in the target string, so the pattern C<blurfl> |
---|
929 | would match "blurfl" in the target string. |
---|
930 | |
---|
931 | You can specify a character class, by enclosing a list of characters |
---|
932 | in C<[]>, which will match any one character from the list. If the |
---|
933 | first character after the "[" is "^", the class matches any character not |
---|
934 | in the list. Within a list, the "-" character specifies a |
---|
935 | range, so that C<a-z> represents all characters between "a" and "z", |
---|
936 | inclusive. If you want either "-" or "]" itself to be a member of a |
---|
937 | class, put it at the start of the list (possibly after a "^"), or |
---|
938 | escape it with a backslash. "-" is also taken literally when it is |
---|
939 | at the end of the list, just before the closing "]". (The |
---|
940 | following all specify the same class of three characters: C<[-az]>, |
---|
941 | C<[az-]>, and C<[a\-z]>. All are different from C<[a-z]>, which |
---|
942 | specifies a class containing twenty-six characters.) |
---|
943 | Also, if you try to use the character classes C<\w>, C<\W>, C<\s>, |
---|
944 | C<\S>, C<\d>, or C<\D> as endpoints of a range, that's not a range, |
---|
945 | the "-" is understood literally. |
---|
946 | |
---|
947 | Note also that the whole range idea is rather unportable between |
---|
948 | character sets--and even within character sets they may cause results |
---|
949 | you probably didn't expect. A sound principle is to use only ranges |
---|
950 | that begin from and end at either alphabets of equal case ([a-e], |
---|
951 | [A-E]), or digits ([0-9]). Anything else is unsafe. If in doubt, |
---|
952 | spell out the character sets in full. |
---|
953 | |
---|
954 | Characters may be specified using a metacharacter syntax much like that |
---|
955 | used in C: "\n" matches a newline, "\t" a tab, "\r" a carriage return, |
---|
956 | "\f" a form feed, etc. More generally, \I<nnn>, where I<nnn> is a string |
---|
957 | of octal digits, matches the character whose ASCII value is I<nnn>. |
---|
958 | Similarly, \xI<nn>, where I<nn> are hexadecimal digits, matches the |
---|
959 | character whose ASCII value is I<nn>. The expression \cI<x> matches the |
---|
960 | ASCII character control-I<x>. Finally, the "." metacharacter matches any |
---|
961 | character except "\n" (unless you use C</s>). |
---|
962 | |
---|
963 | You can specify a series of alternatives for a pattern using "|" to |
---|
964 | separate them, so that C<fee|fie|foe> will match any of "fee", "fie", |
---|
965 | or "foe" in the target string (as would C<f(e|i|o)e>). The |
---|
966 | first alternative includes everything from the last pattern delimiter |
---|
967 | ("(", "[", or the beginning of the pattern) up to the first "|", and |
---|
968 | the last alternative contains everything from the last "|" to the next |
---|
969 | pattern delimiter. That's why it's common practice to include |
---|
970 | alternatives in parentheses: to minimize confusion about where they |
---|
971 | start and end. |
---|
972 | |
---|
973 | Alternatives are tried from left to right, so the first |
---|
974 | alternative found for which the entire expression matches, is the one that |
---|
975 | is chosen. This means that alternatives are not necessarily greedy. For |
---|
976 | example: when matching C<foo|foot> against "barefoot", only the "foo" |
---|
977 | part will match, as that is the first alternative tried, and it successfully |
---|
978 | matches the target string. (This might not seem important, but it is |
---|
979 | important when you are capturing matched text using parentheses.) |
---|
980 | |
---|
981 | Also remember that "|" is interpreted as a literal within square brackets, |
---|
982 | so if you write C<[fee|fie|foe]> you're really only matching C<[feio|]>. |
---|
983 | |
---|
984 | Within a pattern, you may designate subpatterns for later reference |
---|
985 | by enclosing them in parentheses, and you may refer back to the |
---|
986 | I<n>th subpattern later in the pattern using the metacharacter |
---|
987 | \I<n>. Subpatterns are numbered based on the left to right order |
---|
988 | of their opening parenthesis. A backreference matches whatever |
---|
989 | actually matched the subpattern in the string being examined, not |
---|
990 | the rules for that subpattern. Therefore, C<(0|0x)\d*\s\1\d*> will |
---|
991 | match "0x1234 0x4321", but not "0x1234 01234", because subpattern |
---|
992 | 1 matched "0x", even though the rule C<0|0x> could potentially match |
---|
993 | the leading 0 in the second number. |
---|
994 | |
---|
995 | =head2 Warning on \1 vs $1 |
---|
996 | |
---|
997 | Some people get too used to writing things like: |
---|
998 | |
---|
999 | $pattern =~ s/(\W)/\\\1/g; |
---|
1000 | |
---|
1001 | This is grandfathered for the RHS of a substitute to avoid shocking the |
---|
1002 | B<sed> addicts, but it's a dirty habit to get into. That's because in |
---|
1003 | PerlThink, the righthand side of a C<s///> is a double-quoted string. C<\1> in |
---|
1004 | the usual double-quoted string means a control-A. The customary Unix |
---|
1005 | meaning of C<\1> is kludged in for C<s///>. However, if you get into the habit |
---|
1006 | of doing that, you get yourself into trouble if you then add an C</e> |
---|
1007 | modifier. |
---|
1008 | |
---|
1009 | s/(\d+)/ \1 + 1 /eg; # causes warning under -w |
---|
1010 | |
---|
1011 | Or if you try to do |
---|
1012 | |
---|
1013 | s/(\d+)/\1000/; |
---|
1014 | |
---|
1015 | You can't disambiguate that by saying C<\{1}000>, whereas you can fix it with |
---|
1016 | C<${1}000>. The operation of interpolation should not be confused |
---|
1017 | with the operation of matching a backreference. Certainly they mean two |
---|
1018 | different things on the I<left> side of the C<s///>. |
---|
1019 | |
---|
1020 | =head2 Repeated patterns matching zero-length substring |
---|
1021 | |
---|
1022 | B<WARNING>: Difficult material (and prose) ahead. This section needs a rewrite. |
---|
1023 | |
---|
1024 | Regular expressions provide a terse and powerful programming language. As |
---|
1025 | with most other power tools, power comes together with the ability |
---|
1026 | to wreak havoc. |
---|
1027 | |
---|
1028 | A common abuse of this power stems from the ability to make infinite |
---|
1029 | loops using regular expressions, with something as innocuous as: |
---|
1030 | |
---|
1031 | 'foo' =~ m{ ( o? )* }x; |
---|
1032 | |
---|
1033 | The C<o?> can match at the beginning of C<'foo'>, and since the position |
---|
1034 | in the string is not moved by the match, C<o?> would match again and again |
---|
1035 | because of the C<*> modifier. Another common way to create a similar cycle |
---|
1036 | is with the looping modifier C<//g>: |
---|
1037 | |
---|
1038 | @matches = ( 'foo' =~ m{ o? }xg ); |
---|
1039 | |
---|
1040 | or |
---|
1041 | |
---|
1042 | print "match: <$&>\n" while 'foo' =~ m{ o? }xg; |
---|
1043 | |
---|
1044 | or the loop implied by split(). |
---|
1045 | |
---|
1046 | However, long experience has shown that many programming tasks may |
---|
1047 | be significantly simplified by using repeated subexpressions that |
---|
1048 | may match zero-length substrings. Here's a simple example being: |
---|
1049 | |
---|
1050 | @chars = split //, $string; # // is not magic in split |
---|
1051 | ($whitewashed = $string) =~ s/()/ /g; # parens avoid magic s// / |
---|
1052 | |
---|
1053 | Thus Perl allows such constructs, by I<forcefully breaking |
---|
1054 | the infinite loop>. The rules for this are different for lower-level |
---|
1055 | loops given by the greedy modifiers C<*+{}>, and for higher-level |
---|
1056 | ones like the C</g> modifier or split() operator. |
---|
1057 | |
---|
1058 | The lower-level loops are I<interrupted> (that is, the loop is |
---|
1059 | broken) when Perl detects that a repeated expression matched a |
---|
1060 | zero-length substring. Thus |
---|
1061 | |
---|
1062 | m{ (?: NON_ZERO_LENGTH | ZERO_LENGTH )* }x; |
---|
1063 | |
---|
1064 | is made equivalent to |
---|
1065 | |
---|
1066 | m{ (?: NON_ZERO_LENGTH )* |
---|
1067 | | |
---|
1068 | (?: ZERO_LENGTH )? |
---|
1069 | }x; |
---|
1070 | |
---|
1071 | The higher level-loops preserve an additional state between iterations: |
---|
1072 | whether the last match was zero-length. To break the loop, the following |
---|
1073 | match after a zero-length match is prohibited to have a length of zero. |
---|
1074 | This prohibition interacts with backtracking (see L<"Backtracking">), |
---|
1075 | and so the I<second best> match is chosen if the I<best> match is of |
---|
1076 | zero length. |
---|
1077 | |
---|
1078 | For example: |
---|
1079 | |
---|
1080 | $_ = 'bar'; |
---|
1081 | s/\w??/<$&>/g; |
---|
1082 | |
---|
1083 | results in C<"<><b><><a><><r><>">. At each position of the string the best |
---|
1084 | match given by non-greedy C<??> is the zero-length match, and the I<second |
---|
1085 | best> match is what is matched by C<\w>. Thus zero-length matches |
---|
1086 | alternate with one-character-long matches. |
---|
1087 | |
---|
1088 | Similarly, for repeated C<m/()/g> the second-best match is the match at the |
---|
1089 | position one notch further in the string. |
---|
1090 | |
---|
1091 | The additional state of being I<matched with zero-length> is associated with |
---|
1092 | the matched string, and is reset by each assignment to pos(). |
---|
1093 | Zero-length matches at the end of the previous match are ignored |
---|
1094 | during C<split>. |
---|
1095 | |
---|
1096 | =head2 Combining pieces together |
---|
1097 | |
---|
1098 | Each of the elementary pieces of regular expressions which were described |
---|
1099 | before (such as C<ab> or C<\Z>) could match at most one substring |
---|
1100 | at the given position of the input string. However, in a typical regular |
---|
1101 | expression these elementary pieces are combined into more complicated |
---|
1102 | patterns using combining operators C<ST>, C<S|T>, C<S*> etc |
---|
1103 | (in these examples C<S> and C<T> are regular subexpressions). |
---|
1104 | |
---|
1105 | Such combinations can include alternatives, leading to a problem of choice: |
---|
1106 | if we match a regular expression C<a|ab> against C<"abc">, will it match |
---|
1107 | substring C<"a"> or C<"ab">? One way to describe which substring is |
---|
1108 | actually matched is the concept of backtracking (see L<"Backtracking">). |
---|
1109 | However, this description is too low-level and makes you think |
---|
1110 | in terms of a particular implementation. |
---|
1111 | |
---|
1112 | Another description starts with notions of "better"/"worse". All the |
---|
1113 | substrings which may be matched by the given regular expression can be |
---|
1114 | sorted from the "best" match to the "worst" match, and it is the "best" |
---|
1115 | match which is chosen. This substitutes the question of "what is chosen?" |
---|
1116 | by the question of "which matches are better, and which are worse?". |
---|
1117 | |
---|
1118 | Again, for elementary pieces there is no such question, since at most |
---|
1119 | one match at a given position is possible. This section describes the |
---|
1120 | notion of better/worse for combining operators. In the description |
---|
1121 | below C<S> and C<T> are regular subexpressions. |
---|
1122 | |
---|
1123 | =over |
---|
1124 | |
---|
1125 | =item C<ST> |
---|
1126 | |
---|
1127 | Consider two possible matches, C<AB> and C<A'B'>, C<A> and C<A'> are |
---|
1128 | substrings which can be matched by C<S>, C<B> and C<B'> are substrings |
---|
1129 | which can be matched by C<T>. |
---|
1130 | |
---|
1131 | If C<A> is better match for C<S> than C<A'>, C<AB> is a better |
---|
1132 | match than C<A'B'>. |
---|
1133 | |
---|
1134 | If C<A> and C<A'> coincide: C<AB> is a better match than C<AB'> if |
---|
1135 | C<B> is better match for C<T> than C<B'>. |
---|
1136 | |
---|
1137 | =item C<S|T> |
---|
1138 | |
---|
1139 | When C<S> can match, it is a better match than when only C<T> can match. |
---|
1140 | |
---|
1141 | Ordering of two matches for C<S> is the same as for C<S>. Similar for |
---|
1142 | two matches for C<T>. |
---|
1143 | |
---|
1144 | =item C<S{REPEAT_COUNT}> |
---|
1145 | |
---|
1146 | Matches as C<SSS...S> (repeated as many times as necessary). |
---|
1147 | |
---|
1148 | =item C<S{min,max}> |
---|
1149 | |
---|
1150 | Matches as C<S{max}|S{max-1}|...|S{min+1}|S{min}>. |
---|
1151 | |
---|
1152 | =item C<S{min,max}?> |
---|
1153 | |
---|
1154 | Matches as C<S{min}|S{min+1}|...|S{max-1}|S{max}>. |
---|
1155 | |
---|
1156 | =item C<S?>, C<S*>, C<S+> |
---|
1157 | |
---|
1158 | Same as C<S{0,1}>, C<S{0,BIG_NUMBER}>, C<S{1,BIG_NUMBER}> respectively. |
---|
1159 | |
---|
1160 | =item C<S??>, C<S*?>, C<S+?> |
---|
1161 | |
---|
1162 | Same as C<S{0,1}?>, C<S{0,BIG_NUMBER}?>, C<S{1,BIG_NUMBER}?> respectively. |
---|
1163 | |
---|
1164 | =item C<< (?>S) >> |
---|
1165 | |
---|
1166 | Matches the best match for C<S> and only that. |
---|
1167 | |
---|
1168 | =item C<(?=S)>, C<(?<=S)> |
---|
1169 | |
---|
1170 | Only the best match for C<S> is considered. (This is important only if |
---|
1171 | C<S> has capturing parentheses, and backreferences are used somewhere |
---|
1172 | else in the whole regular expression.) |
---|
1173 | |
---|
1174 | =item C<(?!S)>, C<(?<!S)> |
---|
1175 | |
---|
1176 | For this grouping operator there is no need to describe the ordering, since |
---|
1177 | only whether or not C<S> can match is important. |
---|
1178 | |
---|
1179 | =item C<(??{ EXPR })> |
---|
1180 | |
---|
1181 | The ordering is the same as for the regular expression which is |
---|
1182 | the result of EXPR. |
---|
1183 | |
---|
1184 | =item C<(?(condition)yes-pattern|no-pattern)> |
---|
1185 | |
---|
1186 | Recall that which of C<yes-pattern> or C<no-pattern> actually matches is |
---|
1187 | already determined. The ordering of the matches is the same as for the |
---|
1188 | chosen subexpression. |
---|
1189 | |
---|
1190 | =back |
---|
1191 | |
---|
1192 | The above recipes describe the ordering of matches I<at a given position>. |
---|
1193 | One more rule is needed to understand how a match is determined for the |
---|
1194 | whole regular expression: a match at an earlier position is always better |
---|
1195 | than a match at a later position. |
---|
1196 | |
---|
1197 | =head2 Creating custom RE engines |
---|
1198 | |
---|
1199 | Overloaded constants (see L<overload>) provide a simple way to extend |
---|
1200 | the functionality of the RE engine. |
---|
1201 | |
---|
1202 | Suppose that we want to enable a new RE escape-sequence C<\Y|> which |
---|
1203 | matches at boundary between white-space characters and non-whitespace |
---|
1204 | characters. Note that C<(?=\S)(?<!\S)|(?!\S)(?<=\S)> matches exactly |
---|
1205 | at these positions, so we want to have each C<\Y|> in the place of the |
---|
1206 | more complicated version. We can create a module C<customre> to do |
---|
1207 | this: |
---|
1208 | |
---|
1209 | package customre; |
---|
1210 | use overload; |
---|
1211 | |
---|
1212 | sub import { |
---|
1213 | shift; |
---|
1214 | die "No argument to customre::import allowed" if @_; |
---|
1215 | overload::constant 'qr' => \&convert; |
---|
1216 | } |
---|
1217 | |
---|
1218 | sub invalid { die "/$_[0]/: invalid escape '\\$_[1]'"} |
---|
1219 | |
---|
1220 | my %rules = ( '\\' => '\\', |
---|
1221 | 'Y|' => qr/(?=\S)(?<!\S)|(?!\S)(?<=\S)/ ); |
---|
1222 | sub convert { |
---|
1223 | my $re = shift; |
---|
1224 | $re =~ s{ |
---|
1225 | \\ ( \\ | Y . ) |
---|
1226 | } |
---|
1227 | { $rules{$1} or invalid($re,$1) }sgex; |
---|
1228 | return $re; |
---|
1229 | } |
---|
1230 | |
---|
1231 | Now C<use customre> enables the new escape in constant regular |
---|
1232 | expressions, i.e., those without any runtime variable interpolations. |
---|
1233 | As documented in L<overload>, this conversion will work only over |
---|
1234 | literal parts of regular expressions. For C<\Y|$re\Y|> the variable |
---|
1235 | part of this regular expression needs to be converted explicitly |
---|
1236 | (but only if the special meaning of C<\Y|> should be enabled inside $re): |
---|
1237 | |
---|
1238 | use customre; |
---|
1239 | $re = <>; |
---|
1240 | chomp $re; |
---|
1241 | $re = customre::convert $re; |
---|
1242 | /\Y|$re\Y|/; |
---|
1243 | |
---|
1244 | =head1 BUGS |
---|
1245 | |
---|
1246 | This document varies from difficult to understand to completely |
---|
1247 | and utterly opaque. The wandering prose riddled with jargon is |
---|
1248 | hard to fathom in several places. |
---|
1249 | |
---|
1250 | This document needs a rewrite that separates the tutorial content |
---|
1251 | from the reference content. |
---|
1252 | |
---|
1253 | =head1 SEE ALSO |
---|
1254 | |
---|
1255 | L<perlop/"Regexp Quote-Like Operators">. |
---|
1256 | |
---|
1257 | L<perlop/"Gory details of parsing quoted constructs">. |
---|
1258 | |
---|
1259 | L<perlfaq6>. |
---|
1260 | |
---|
1261 | L<perlfunc/pos>. |
---|
1262 | |
---|
1263 | L<perllocale>. |
---|
1264 | |
---|
1265 | I<Mastering Regular Expressions> by Jeffrey Friedl, published |
---|
1266 | by O'Reilly and Associates. |
---|