This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
PATCH: [perl #128219] typo in perlrecharclass
[perl5.git] / pod / perlrecharclass.pod
CommitLineData
8a118206 1=head1 NAME
ea449505 2X<character class>
8a118206
RGS
3
4perlrecharclass - Perl Regular Expression Character Classes
5
6=head1 DESCRIPTION
7
8The top level documentation about Perl regular expressions
9is found in L<perlre>.
10
11This manual page discusses the syntax and use of character
6b83a163 12classes in Perl regular expressions.
8a118206 13
6b83a163 14A character class is a way of denoting a set of characters
8a118206 15in such a way that one character of the set is matched.
6b83a163 16It's important to remember that: matching a character class
8a118206
RGS
17consumes exactly one character in the source string. (The source
18string is the string the regular expression is matched against.)
19
20There are three types of character classes in Perl regular
6b83a163 21expressions: the dot, backslash sequences, and the form enclosed in square
ea449505 22brackets. Keep in mind, though, that often the term "character class" is used
6b83a163 23to mean just the bracketed form. Certainly, most Perl documentation does that.
8a118206
RGS
24
25=head2 The dot
26
27The dot (or period), C<.> is probably the most used, and certainly
28the most well-known character class. By default, a dot matches any
5db9882c 29character, except for the newline. That default can be changed to
6b83a163
KW
30add matching the newline by using the I<single line> modifier: either
31for the entire regular expression with the C</s> modifier, or
d66e1f56 32locally with C<(?s)>. (The C<L</\N>> backslash sequence, described
6b83a163
KW
33below, matches any character except newline without regard to the
34I<single line> modifier.)
8a118206
RGS
35
36Here are some examples:
37
38 "a" =~ /./ # Match
39 "." =~ /./ # Match
40 "" =~ /./ # No match (dot has to match a character)
41 "\n" =~ /./ # No match (dot does not match a newline)
42 "\n" =~ /./s # Match (global 'single line' modifier)
43 "\n" =~ /(?s:.)/ # Match (local 'single line' modifier)
44 "ab" =~ /^.$/ # No match (dot matches one character)
45
6b83a163 46=head2 Backslash sequences
82206b5e 47X<\w> X<\W> X<\s> X<\S> X<\d> X<\D> X<\p> X<\P>
ea449505
KW
48X<\N> X<\v> X<\V> X<\h> X<\H>
49X<word> X<whitespace>
8a118206 50
6b83a163
KW
51A backslash sequence is a sequence of characters, the first one of which is a
52backslash. Perl ascribes special meaning to many such sequences, and some of
53these are character classes. That is, they match a single character each,
54provided that the character belongs to the specific set of characters defined
55by the sequence.
8a118206 56
6b83a163
KW
57Here's a list of the backslash sequences that are character classes. They
58are discussed in more detail below. (For the backslash sequences that aren't
59character classes, see L<perlrebackslash>.)
8a118206 60
6b83a163
KW
61 \d Match a decimal digit character.
62 \D Match a non-decimal-digit character.
8a118206
RGS
63 \w Match a "word" character.
64 \W Match a non-"word" character.
ea449505
KW
65 \s Match a whitespace character.
66 \S Match a non-whitespace character.
67 \h Match a horizontal whitespace character.
68 \H Match a character that isn't horizontal whitespace.
ea449505
KW
69 \v Match a vertical whitespace character.
70 \V Match a character that isn't vertical whitespace.
4e5e0888 71 \N Match a character that isn't a newline.
6b83a163 72 \pP, \p{Prop} Match a character that has the given Unicode property.
6c5a041f 73 \PP, \P{Prop} Match a character that doesn't have the Unicode property
8a118206 74
1433f837
KW
75=head3 \N
76
2171640d 77C<\N>, available starting in v5.12, like the dot, matches any
1433f837
KW
78character that is not a newline. The difference is that C<\N> is not influenced
79by the I<single line> regular expression modifier (see L</The dot> above). Note
80that the form C<\N{...}> may mean something completely different. When the
81C<{...}> is a L<quantifier|perlre/Quantifiers>, it means to match a non-newline
82character that many times. For example, C<\N{3}> means to match 3
83non-newlines; C<\N{5,}> means to match 5 or more non-newlines. But if C<{...}>
84is not a legal quantifier, it is presumed to be a named character. See
85L<charnames> for those. For example, none of C<\N{COLON}>, C<\N{4F}>, and
86C<\N{F4}> contain legal quantifiers, so Perl will try to find characters whose
87names are respectively C<COLON>, C<4F>, and C<F4>.
88
8a118206
RGS
89=head3 Digits
90
b6538e4f 91C<\d> matches a single character considered to be a decimal I<digit>.
5db9882c 92If the C</a> regular expression modifier is in effect, it matches [0-9].
582da942 93Otherwise, it
82206b5e
KW
94matches anything that is matched by C<\p{Digit}>, which includes [0-9].
95(An unlikely possible exception is that under locale matching rules, the
d66e1f56
KW
96current locale might not have C<[0-9]> matched by C<\d>, and/or might match
97other characters whose code point is less than 256. The only such locale
98definitions that are legal would be to match C<[0-9]> plus another set of
9910 consecutive digit characters; anything else would be in violation of
100the C language standard, but Perl doesn't currently assume anything in
101regard to this.)
82206b5e
KW
102
103What this means is that unless the C</a> modifier is in effect C<\d> not
104only matches the digits '0' - '9', but also Arabic, Devanagari, and
105digits from other languages. This may cause some confusion, and some
106security issues.
107
108Some digits that C<\d> matches look like some of the [0-9] ones, but
109have different values. For example, BENGALI DIGIT FOUR (U+09EA) looks
110very much like an ASCII DIGIT EIGHT (U+0038). An application that
111is expecting only the ASCII digits might be misled, or if the match is
112C<\d+>, the matched string might contain a mixture of digits from
113different writing systems that look like they signify a number different
67592e11 114than they actually do. L<Unicode::UCD/num()> can
e397bccf 115be used to safely
82206b5e
KW
116calculate the value, returning C<undef> if the input string contains
117such a mixture.
118
119What C<\p{Digit}> means (and hence C<\d> except under the C</a>
120modifier) is C<\p{General_Category=Decimal_Number}>, or synonymously,
121C<\p{General_Category=Digit}>. Starting with Unicode version 4.1, this
122is the same set of characters matched by C<\p{Numeric_Type=Decimal}>.
6b83a163
KW
123But Unicode also has a different property with a similar name,
124C<\p{Numeric_Type=Digit}>, which matches a completely different set of
82206b5e
KW
125characters. These characters are things such as C<CIRCLED DIGIT ONE>
126or subscripts, or are from writing systems that lack all ten digits.
6b83a163 127
82206b5e
KW
128The design intent is for C<\d> to exactly match the set of characters
129that can safely be used with "normal" big-endian positional decimal
130syntax, where, for example 123 means one 'hundred', plus two 'tens',
131plus three 'ones'. This positional notation does not necessarily apply
132to characters that match the other type of "digit",
133C<\p{Numeric_Type=Digit}>, and so C<\d> doesn't match them.
6b83a163 134
e2cfb18c 135The Tamil digits (U+0BE6 - U+0BEF) can also legally be
82206b5e
KW
136used in old-style Tamil numbers in which they would appear no more than
137one in a row, separated by characters that mean "times 10", "times 100",
138etc. (See L<http://www.unicode.org/notes/tn21>.)
8a118206 139
b6538e4f 140Any character not matched by C<\d> is matched by C<\D>.
8a118206
RGS
141
142=head3 Word characters
143
ea449505 144A C<\w> matches a single alphanumeric character (an alphabetic character, or a
41805eb9
KW
145decimal digit); or a connecting punctuation character, such as an
146underscore ("_"); or a "mark" character (like some sort of accent) that
147attaches to one of those. It does not match a whole word. To match a
148whole word, use C<\w+>. This isn't the same thing as matching an
149English word, but in the ASCII range it is the same as a string of
150Perl-identifier characters.
82206b5e
KW
151
152=over
153
154=item If the C</a> modifier is in effect ...
155
156C<\w> matches the 63 characters [a-zA-Z0-9_].
157
158=item otherwise ...
159
160=over
161
162=item For code points above 255 ...
163
164C<\w> matches the same as C<\p{Word}> matches in this range. That is,
165it matches Thai letters, Greek letters, etc. This includes connector
d35dd6c6 166punctuation (like the underscore) which connect two words together, or
b6538e4f 167diacritics, such as a C<COMBINING TILDE> and the modifier letters, which
82206b5e
KW
168are generally used to add auxiliary markings to letters.
169
170=item For code points below 256 ...
171
172=over
173
174=item if locale rules are in effect ...
175
176C<\w> matches the platform's native underscore character plus whatever
177the locale considers to be alphanumeric.
178
4b9734bf 179=item if Unicode rules are in effect ...
82206b5e
KW
180
181C<\w> matches exactly what C<\p{Word}> matches.
182
183=item otherwise ...
184
185C<\w> matches [a-zA-Z0-9_].
186
187=back
188
189=back
190
191=back
192
193Which rules apply are determined as described in L<perlre/Which character set modifier is in effect?>.
8a118206 194
6b83a163
KW
195There are a number of security issues with the full Unicode list of word
196characters. See L<http://unicode.org/reports/tr36>.
197
198Also, for a somewhat finer-grained set of characters that are in programming
199language identifiers beyond the ASCII range, you may wish to instead use the
e2cfb18c
KW
200more customized L</Unicode Properties>, C<\p{ID_Start}>,
201C<\p{ID_Continue}>, C<\p{XID_Start}>, and C<\p{XID_Continue}>. See
202L<http://unicode.org/reports/tr31>.
6b83a163 203
b6538e4f 204Any character not matched by C<\w> is matched by C<\W>.
8a118206 205
ea449505
KW
206=head3 Whitespace
207
82206b5e
KW
208C<\s> matches any single character considered whitespace.
209
210=over
211
212=item If the C</a> modifier is in effect ...
213
d28d8023
KW
214In all Perl versions, C<\s> matches the 5 characters [\t\n\f\r ]; that
215is, the horizontal tab,
216the newline, the form feed, the carriage return, and the space.
779cf272 217Starting in Perl v5.18, it also matches the vertical tab, C<\cK>.
d28d8023 218See note C<[1]> below for a discussion of this.
82206b5e
KW
219
220=item otherwise ...
221
222=over
223
224=item For code points above 255 ...
225
226C<\s> matches exactly the code points above 255 shown with an "s" column
227in the table below.
228
229=item For code points below 256 ...
230
231=over
232
233=item if locale rules are in effect ...
234
d28d8023 235C<\s> matches whatever the locale considers to be whitespace.
82206b5e 236
4b9734bf 237=item if Unicode rules are in effect ...
82206b5e
KW
238
239C<\s> matches exactly the characters shown with an "s" column in the
240table below.
241
242=item otherwise ...
243
779cf272 244C<\s> matches [\t\n\f\r ] and, starting in Perl
d28d8023
KW
245v5.18, the vertical tab, C<\cK>.
246(See note C<[1]> below for a discussion of this.)
82206b5e
KW
247Note that this list doesn't include the non-breaking space.
248
249=back
250
251=back
252
253=back
254
255Which rules apply are determined as described in L<perlre/Which character set modifier is in effect?>.
8a118206 256
b6538e4f 257Any character not matched by C<\s> is matched by C<\S>.
8a118206 258
b6538e4f 259C<\h> matches any character considered horizontal whitespace;
8129baca 260this includes the platform's space and tab characters and several others
b6538e4f 261listed in the table below. C<\H> matches any character
8129baca
KW
262not considered horizontal whitespace. They use the platform's native
263character set, and do not consider any locale that may otherwise be in
264use.
ea449505 265
b6538e4f 266C<\v> matches any character considered vertical whitespace;
8129baca 267this includes the platform's carriage return and line feed characters (newline)
b6538e4f
TC
268plus several other characters, all listed in the table below.
269C<\V> matches any character not considered vertical whitespace.
8129baca
KW
270They use the platform's native character set, and do not consider any
271locale that may otherwise be in use.
8a118206
RGS
272
273C<\R> matches anything that can be considered a newline under Unicode
412a49a2
KW
274rules. It can match a multi-character sequence. It cannot be used inside
275a bracketed character class; use C<\v> instead (vertical whitespace).
276It uses the platform's
8129baca
KW
277native character set, and does not consider any locale that may
278otherwise be in use.
ea449505 279Details are discussed in L<perlrebackslash>.
8a118206 280
82206b5e 281Note that unlike C<\s> (and C<\d> and C<\w>), C<\h> and C<\v> always match
8129baca
KW
282the same characters, without regard to other factors, such as the active
283locale or whether the source string is in UTF-8 format.
8a118206 284
d28d8023
KW
285One might think that C<\s> is equivalent to C<[\h\v]>. This is indeed true
286starting in Perl v5.18, but prior to that, the sole difference was that the
287vertical tab (C<"\cK">) was not matched by C<\s>.
8a118206
RGS
288
289The following table is a complete listing of characters matched by
a9c9e371 290C<\s>, C<\h> and C<\v> as of Unicode 6.3.
8a118206 291
582da942 292The first column gives the Unicode code point of the character (in hex format),
8a118206 293the second column gives the (Unicode) name. The third column indicates
4b9734bf
KW
294by which class(es) the character is matched (assuming no locale is in
295effect that changes the C<\s> matching).
8a118206 296
fc28d2a3
KW
297 0x0009 CHARACTER TABULATION h s
298 0x000a LINE FEED (LF) vs
d28d8023 299 0x000b LINE TABULATION vs [1]
fc28d2a3
KW
300 0x000c FORM FEED (FF) vs
301 0x000d CARRIAGE RETURN (CR) vs
302 0x0020 SPACE h s
d28d8023
KW
303 0x0085 NEXT LINE (NEL) vs [2]
304 0x00a0 NO-BREAK SPACE h s [2]
fc28d2a3 305 0x1680 OGHAM SPACE MARK h s
fc28d2a3
KW
306 0x2000 EN QUAD h s
307 0x2001 EM QUAD h s
308 0x2002 EN SPACE h s
309 0x2003 EM SPACE h s
310 0x2004 THREE-PER-EM SPACE h s
311 0x2005 FOUR-PER-EM SPACE h s
312 0x2006 SIX-PER-EM SPACE h s
313 0x2007 FIGURE SPACE h s
314 0x2008 PUNCTUATION SPACE h s
315 0x2009 THIN SPACE h s
316 0x200a HAIR SPACE h s
317 0x2028 LINE SEPARATOR vs
318 0x2029 PARAGRAPH SEPARATOR vs
319 0x202f NARROW NO-BREAK SPACE h s
320 0x205f MEDIUM MATHEMATICAL SPACE h s
321 0x3000 IDEOGRAPHIC SPACE h s
8a118206
RGS
322
323=over 4
324
325=item [1]
326
779cf272
KW
327Prior to Perl v5.18, C<\s> did not match the vertical tab.
328C<[^\S\cK]> (obscurely) matches what C<\s> traditionally did.
d28d8023
KW
329
330=item [2]
331
82206b5e
KW
332NEXT LINE and NO-BREAK SPACE may or may not match C<\s> depending
333on the rules in effect. See
334L<the beginning of this section|/Whitespace>.
8a118206
RGS
335
336=back
337
8a118206
RGS
338=head3 Unicode Properties
339
c1c4ae3a
KW
340C<\pP> and C<\p{Prop}> are character classes to match characters that fit given
341Unicode properties. One letter property names can be used in the C<\pP> form,
342with the property name following the C<\p>, otherwise, braces are required.
343When using braces, there is a single form, which is just the property name
344enclosed in the braces, and a compound form which looks like C<\p{name=value}>,
b6538e4f 345which means to match if the property "name" for the character has that particular
c1c4ae3a 346"value".
e1b711da
KW
347For instance, a match for a number can be written as C</\pN/> or as
348C</\p{Number}/>, or as C</\p{Number=True}/>.
349Lowercase letters are matched by the property I<Lowercase_Letter> which
e2cfb18c 350has the short form I<Ll>. They need the braces, so are written as C</\p{Ll}/> or
e1b711da
KW
351C</\p{Lowercase_Letter}/>, or C</\p{General_Category=Lowercase_Letter}/>
352(the underscores are optional).
353C</\pLl/> is valid, but means something different.
8a118206
RGS
354It matches a two character string: a letter (Unicode property C<\pL>),
355followed by a lowercase C<l>.
356
bc943be5 357If locale rules are not in effect, the use of
82206b5e 358a Unicode property will force the regular expression into using Unicode
bc943be5 359rules, if it isn't already.
82206b5e 360
56ca34ca
KW
361Note that almost all properties are immune to case-insensitive matching.
362That is, adding a C</i> regular expression modifier does not change what
82206b5e 363they match. There are two sets that are affected. The first set is
56ca34ca
KW
364C<Uppercase_Letter>,
365C<Lowercase_Letter>,
366and C<Titlecase_Letter>,
367all of which match C<Cased_Letter> under C</i> matching.
b6538e4f 368The second set is
56ca34ca
KW
369C<Uppercase>,
370C<Lowercase>,
371and C<Titlecase>,
372all of which match C<Cased> under C</i> matching.
373(The difference between these sets is that some things, such as Roman
e2cfb18c 374numerals, come in both upper and lower case, so they are C<Cased>, but
b6538e4f 375aren't considered to be letters, so they aren't C<Cased_Letter>s. They're
82206b5e
KW
376actually C<Letter_Number>s.)
377This set also includes its subsets C<PosixUpper> and C<PosixLower>, both
e2cfb18c 378of which under C</i> match C<PosixAlpha>.
56ca34ca
KW
379
380For more details on Unicode properties, see L<perlunicode/Unicode
381Character Properties>; for a
e1b711da 382complete list of possible properties, see
56ca34ca
KW
383L<perluniprops/Properties accessible through \p{} and \P{}>,
384which notes all forms that have C</i> differences.
e1b711da 385It is also possible to define your own properties. This is discussed in
8a118206
RGS
386L<perlunicode/User-Defined Character Properties>.
387
94b42e47 388Unicode properties are defined (surprise!) only on Unicode code points.
2d88a86a
KW
389Starting in v5.20, when matching against C<\p> and C<\P>, Perl treats
390non-Unicode code points (those above the legal Unicode maximum of
3910x10FFFF) as if they were typical unassigned Unicode code points.
94b42e47 392
2d88a86a
KW
393Prior to v5.20, Perl raised a warning and made all matches fail on
394non-Unicode code points. This could be somewhat surprising:
94b42e47 395
2d88a86a
KW
396 chr(0x110000) =~ \p{ASCII_Hex_Digit=True} # Fails on Perls < v5.20.
397 chr(0x110000) =~ \p{ASCII_Hex_Digit=False} # Also fails on Perls
398 # < v5.20
399
400Even though these two matches might be thought of as complements, until
401v5.20 they were so only on Unicode code points.
94b42e47 402
8a118206
RGS
403=head4 Examples
404
405 "a" =~ /\w/ # Match, "a" is a 'word' character.
406 "7" =~ /\w/ # Match, "7" is a 'word' character as well.
407 "a" =~ /\d/ # No match, "a" isn't a digit.
408 "7" =~ /\d/ # Match, "7" is a digit.
ea449505 409 " " =~ /\s/ # Match, a space is whitespace.
8a118206
RGS
410 "a" =~ /\D/ # Match, "a" is a non-digit.
411 "7" =~ /\D/ # No match, "7" is not a non-digit.
ea449505 412 " " =~ /\S/ # No match, a space is not non-whitespace.
8a118206 413
ea449505
KW
414 " " =~ /\h/ # Match, space is horizontal whitespace.
415 " " =~ /\v/ # No match, space is not vertical whitespace.
416 "\r" =~ /\v/ # Match, a return is vertical whitespace.
8a118206
RGS
417
418 "a" =~ /\pL/ # Match, "a" is a letter.
419 "a" =~ /\p{Lu}/ # No match, /\p{Lu}/ matches upper case letters.
420
421 "\x{0e0b}" =~ /\p{Thai}/ # Match, \x{0e0b} is the character
422 # 'THAI CHARACTER SO SO', and that's in
423 # Thai Unicode class.
ea449505 424 "a" =~ /\P{Lao}/ # Match, as "a" is not a Laotian character.
8a118206 425
82206b5e
KW
426It is worth emphasizing that C<\d>, C<\w>, etc, match single characters, not
427complete numbers or words. To match a number (that consists of digits),
428use C<\d+>; to match a word, use C<\w+>. But be aware of the security
429considerations in doing so, as mentioned above.
8a118206
RGS
430
431=head2 Bracketed Character Classes
432
433The third form of character class you can use in Perl regular expressions
6b83a163 434is the bracketed character class. In its simplest form, it lists the characters
c1c4ae3a 435that may be matched, surrounded by square brackets, like this: C<[aeiou]>.
ea449505 436This matches one of C<a>, C<e>, C<i>, C<o> or C<u>. Like the other
1f59b283 437character classes, exactly one character is matched.* To match
ea449505 438a longer string consisting of characters mentioned in the character
6b83a163 439class, follow the character class with a L<quantifier|perlre/Quantifiers>. For
b6538e4f 440instance, C<[aeiou]+> matches one or more lowercase English vowels.
8a118206
RGS
441
442Repeating a character in a character class has no
443effect; it's considered to be in the set only once.
444
445Examples:
446
447 "e" =~ /[aeiou]/ # Match, as "e" is listed in the class.
448 "p" =~ /[aeiou]/ # No match, "p" is not listed in the class.
449 "ae" =~ /^[aeiou]$/ # No match, a character class only matches
450 # a single character.
451 "ae" =~ /^[aeiou]+$/ # Match, due to the quantifier.
452
1f59b283
KW
453 -------
454
8f0cd35a
KW
455* There are two exceptions to a bracketed character class matching a
456single character only. Each requires special handling by Perl to make
457things work:
458
459=over
460
461=item *
462
463When the class is to match caselessly under C</i> matching rules, and a
464character that is explicitly mentioned inside the class matches a
1f59b283 465multiple-character sequence caselessly under Unicode rules, the class
8f0cd35a
KW
466will also match that sequence. For example, Unicode says that the
467letter C<LATIN SMALL LETTER SHARP S> should match the sequence C<ss>
468under C</i> rules. Thus,
1f59b283
KW
469
470 'ss' =~ /\A\N{LATIN SMALL LETTER SHARP S}\z/i # Matches
471 'ss' =~ /\A[aeioust\N{LATIN SMALL LETTER SHARP S}]\z/i # Matches
472
8f0cd35a
KW
473For this to happen, the class must not be inverted (see L</Negation>)
474and the character must be explicitly specified, and not be part of a
475multi-character range (not even as one of its endpoints). (L</Character
476Ranges> will be explained shortly.) Therefore,
9d53c457 477
eb9e3b14
KW
478 'ss' =~ /\A[\0-\x{ff}]\z/ui # Doesn't match
479 'ss' =~ /\A[\0-\N{LATIN SMALL LETTER SHARP S}]\z/ui # No match
480 'ss' =~ /\A[\xDF-\xDF]\z/ui # Matches on ASCII platforms, since
a845303d 481 # \xDF is LATIN SMALL LETTER SHARP S,
8f0cd35a
KW
482 # and the range is just a single
483 # element
9d53c457
KW
484
485Note that it isn't a good idea to specify these types of ranges anyway.
486
8f0cd35a
KW
487=item *
488
489Some names known to C<\N{...}> refer to a sequence of multiple characters,
490instead of the usual single character. When one of these is included in
491the class, the entire sequence is matched. For example,
492
493 "\N{TAMIL LETTER KA}\N{TAMIL VOWEL SIGN AU}"
494 =~ / ^ [\N{TAMIL SYLLABLE KAU}] $ /x;
495
496matches, because C<\N{TAMIL SYLLABLE KAU}> is a named sequence
497consisting of the two characters matched against. Like the other
eb9e3b14 498instance where a bracketed class can match multiple characters, and for
8f0cd35a
KW
499similar reasons, the class must not be inverted, and the named sequence
500may not appear in a range, even one where it is both endpoints. If
501these happen, it is a fatal error if the character class is within an
502extended L<C<(?[...])>|/Extended Bracketed Character Classes>
503class; and only the first code point is used (with
504a C<regexp>-type warning raised) otherwise.
505
506=back
507
8a118206
RGS
508=head3 Special Characters Inside a Bracketed Character Class
509
510Most characters that are meta characters in regular expressions (that
df225385 511is, characters that carry a special meaning like C<.>, C<*>, or C<(>) lose
8a118206
RGS
512their special meaning and can be used inside a character class without
513the need to escape them. For instance, C<[()]> matches either an opening
514parenthesis, or a closing parenthesis, and the parens inside the character
515class don't group or capture.
516
517Characters that may carry a special meaning inside a character class are:
518C<\>, C<^>, C<->, C<[> and C<]>, and are discussed below. They can be
519escaped with a backslash, although this is sometimes not needed, in which
520case the backslash may be omitted.
521
522The sequence C<\b> is special inside a bracketed character class. While
6b83a163 523outside the character class, C<\b> is an assertion indicating a point
8a118206
RGS
524that does not have either two word characters or two non-word characters
525on either side, inside a bracketed character class, C<\b> matches a
526backspace character.
527
df225385
KW
528The sequences
529C<\a>,
530C<\c>,
531C<\e>,
532C<\f>,
533C<\n>,
e526e8bb 534C<\N{I<NAME>}>,
765fa144 535C<\N{U+I<hex char>}>,
df225385
KW
536C<\r>,
537C<\t>,
538and
539C<\x>
06ee63cd 540are also special and have the same meanings as they do outside a
eb9e3b14 541bracketed character class.
df225385 542
ea449505
KW
543Also, a backslash followed by two or three octal digits is considered an octal
544number.
df225385 545
6b83a163
KW
546A C<[> is not special inside a character class, unless it's the start of a
547POSIX character class (see L</POSIX Character Classes> below). It normally does
548not need escaping.
8a118206 549
6b83a163
KW
550A C<]> is normally either the end of a POSIX character class (see
551L</POSIX Character Classes> below), or it signals the end of the bracketed
552character class. If you want to include a C<]> in the set of characters, you
553must generally escape it.
b6538e4f 554
8a118206
RGS
555However, if the C<]> is the I<first> (or the second if the first
556character is a caret) character of a bracketed character class, it
557does not denote the end of the class (as you cannot have an empty class)
558and is considered part of the set of characters that can be matched without
559escaping.
560
561Examples:
562
563 "+" =~ /[+?*]/ # Match, "+" in a character class is not special.
090752cc 564 "\cH" =~ /[\b]/ # Match, \b inside in a character class
c1c4ae3a 565 # is equivalent to a backspace.
090752cc 566 "]" =~ /[][]/ # Match, as the character class contains
8a118206
RGS
567 # both [ and ].
568 "[]" =~ /[[]]/ # Match, the pattern contains a character class
52f4d632 569 # containing just [, and the character class is
8a118206
RGS
570 # followed by a ].
571
572=head3 Character Ranges
573
574It is not uncommon to want to match a range of characters. Luckily, instead
b6538e4f 575of listing all characters in the range, one may use the hyphen (C<->).
8a118206 576If inside a bracketed character class you have two characters separated
b6538e4f 577by a hyphen, it's treated as if all characters between the two were in
8a118206 578the class. For instance, C<[0-9]> matches any ASCII digit, and C<[a-m]>
e2cfb18c 579matches any lowercase letter from the first half of the ASCII alphabet.
8a118206
RGS
580
581Note that the two characters on either side of the hyphen are not
765fa144 582necessarily both letters or both digits. Any character is possible,
8a118206 583although not advisable. C<['-?]> contains a range of characters, but
b6538e4f 584most people will not know which characters that means. Furthermore,
8a118206
RGS
585such ranges may lead to portability problems if the code has to run on
586a platform that uses a different character set, such as EBCDIC.
587
ea449505
KW
588If a hyphen in a character class cannot syntactically be part of a range, for
589instance because it is the first or the last character of the character class,
b6538e4f
TC
590or if it immediately follows a range, the hyphen isn't special, and so is
591considered a character to be matched literally. If you want a hyphen in
592your set of characters to be matched and its position in the class is such
593that it could be considered part of a range, you must escape that hyphen
594with a backslash.
8a118206
RGS
595
596Examples:
597
598 [a-z] # Matches a character that is a lower case ASCII letter.
c1c4ae3a
KW
599 [a-fz] # Matches any letter between 'a' and 'f' (inclusive) or
600 # the letter 'z'.
8a118206
RGS
601 [-z] # Matches either a hyphen ('-') or the letter 'z'.
602 [a-f-m] # Matches any letter between 'a' and 'f' (inclusive), the
603 # hyphen ('-'), or the letter 'm'.
604 ['-?] # Matches any of the characters '()*+,-./0123456789:;<=>?
605 # (But not on an EBCDIC platform).
c7d25594
KW
606 [\N{APOSTROPHE}-\N{QUESTION MARK}]
607 # Matches any of the characters '()*+,-./0123456789:;<=>?
608 # even on an EBCDIC platform.
ad63362f 609 [\N{U+27}-\N{U+3F}] # Same. (U+27 is "'", and U+3F is "?")
c7d25594
KW
610
611As the final two examples above show, you can achieve portablity to
612non-ASCII platforms by using the C<\N{...}> form for the range
613endpoints. These indicate that the specified range is to be interpreted
614using Unicode values, so C<[\N{U+27}-\N{U+3F}]> means to match
615C<\N{U+27}>, C<\N{U+28}>, C<\N{U+29}>, ..., C<\N{U+3D}>, C<\N{U+3E}>,
616and C<\N{U+3F}>, whatever the native code point versions for those are.
b927b7e9
KW
617These are called "Unicode" ranges. If either end is of the C<\N{...}>
618form, the range is considered Unicode. A C<regexp> warning is raised
619under C<S<"use re 'strict'">> if the other endpoint is specified
620non-portably:
621
622 [\N{U+00}-\x09] # Warning under re 'strict'; \x09 is non-portable
623 [\N{U+00}-\t] # No warning;
624
625Both of the above match the characters C<\N{U+00}> C<\N{U+01}>, ...
626C<\N{U+08}>, C<\N{U+09}>, but the C<\x09> looks like it could be a
627mistake so the warning is raised (under C<re 'strict'>) for it.
c7d25594
KW
628
629Perl also guarantees that the ranges C<A-Z>, C<a-z>, C<0-9>, and any
09e43397 630subranges of these match what an English-only speaker would expect them
c7d25594
KW
631to match on any platform. That is, C<[A-Z]> matches the 26 ASCII
632uppercase letters;
09e43397
KW
633C<[a-z]> matches the 26 lowercase letters; and C<[0-9]> matches the 10
634digits. Subranges, like C<[h-k]>, match correspondingly, in this case
635just the four letters C<"h">, C<"i">, C<"j">, and C<"k">. This is the
636natural behavior on ASCII platforms where the code points (ordinal
637values) for C<"h"> through C<"k"> are consecutive integers (0x68 through
6380x6B). But special handling to achieve this may be needed on platforms
639with a non-ASCII native character set. For example, on EBCDIC
640platforms, the code point for C<"h"> is 0x88, C<"i"> is 0x89, C<"j"> is
6410x91, and C<"k"> is 0x92. Perl specially treats C<[h-k]> to exclude the
642seven code points in the gap: 0x8A through 0x90. This special handling is
643only invoked when the range is a subrange of one of the ASCII uppercase,
644lowercase, and digit ranges, AND each end of the range is expressed
645either as a literal, like C<"A">, or as a named character (C<\N{...}>,
646including the C<\N{U+...> form).
647
648EBCDIC Examples:
649
650 [i-j] # Matches either "i" or "j"
651 [i-\N{LATIN SMALL LETTER J}] # Same
652 [i-\N{U+6A}] # Same
653 [\N{U+69}-\N{U+6A}] # Same
654 [\x{89}-\x{91}] # Matches 0x89 ("i"), 0x8A .. 0x90, 0x91 ("j")
655 [i-\x{91}] # Same
656 [\x{89}-j] # Same
657 [i-J] # Matches, 0x89 ("i") .. 0xC1 ("J"); special
658 # handling doesn't apply because range is mixed
659 # case
8a118206
RGS
660
661=head3 Negation
662
663It is also possible to instead list the characters you do not want to
664match. You can do so by using a caret (C<^>) as the first character in the
b6538e4f 665character class. For instance, C<[^a-z]> matches any character that is not a
e2cfb18c
KW
666lowercase ASCII letter, which therefore includes more than a million
667Unicode code points. The class is said to be "negated" or "inverted".
8a118206
RGS
668
669This syntax make the caret a special character inside a bracketed character
670class, but only if it is the first character of the class. So if you want
82206b5e 671the caret as one of the characters to match, either escape the caret or
e2cfb18c 672else don't list it first.
8a118206 673
1f59b283 674In inverted bracketed character classes, Perl ignores the Unicode rules
8f0cd35a
KW
675that normally say that named sequence, and certain characters should
676match a sequence of multiple characters use under caseless C</i>
677matching. Following those rules could lead to highly confusing
678situations:
1f59b283 679
582da942 680 "ss" =~ /^[^\xDF]+$/ui; # Matches!
1f59b283
KW
681
682This should match any sequences of characters that aren't C<\xDF> nor
683what C<\xDF> matches under C</i>. C<"s"> isn't C<\xDF>, but Unicode
684says that C<"ss"> is what C<\xDF> matches under C</i>. So which one
685"wins"? Do you fail the match because the string has C<ss> or accept it
582da942 686because it has an C<s> followed by another C<s>? Perl has chosen the
8f0cd35a 687latter. (See note in L</Bracketed Character Classes> above.)
1f59b283 688
8a118206
RGS
689Examples:
690
691 "e" =~ /[^aeiou]/ # No match, the 'e' is listed.
692 "x" =~ /[^aeiou]/ # Match, as 'x' isn't a lowercase vowel.
693 "^" =~ /[^^]/ # No match, matches anything that isn't a caret.
694 "^" =~ /[x^]/ # Match, caret is not special here.
695
696=head3 Backslash Sequences
697
ea449505 698You can put any backslash sequence character class (with the exception of
765fa144 699C<\N> and C<\R>) inside a bracketed character class, and it will act just
b6538e4f
TC
700as if you had put all characters matched by the backslash sequence inside the
701character class. For instance, C<[a-f\d]> matches any decimal digit, or any
6b83a163
KW
702of the lowercase letters between 'a' and 'f' inclusive.
703
704C<\N> within a bracketed character class must be of the forms C<\N{I<name>}>
765fa144 705or C<\N{U+I<hex char>}>, and NOT be the form that matches non-newlines,
6b83a163
KW
706for the same reason that a dot C<.> inside a bracketed character class loses
707its special meaning: it matches nearly anything, which generally isn't what you
708want to happen.
df225385 709
8a118206
RGS
710
711Examples:
712
713 /[\p{Thai}\d]/ # Matches a character that is either a Thai
714 # character, or a digit.
715 /[^\p{Arabic}()]/ # Matches a character that is neither an Arabic
716 # character, nor a parenthesis.
717
718Backslash sequence character classes cannot form one of the endpoints
6b83a163
KW
719of a range. Thus, you can't say:
720
721 /[\p{Thai}-\d]/ # Wrong!
8a118206 722
6b83a163 723=head3 POSIX Character Classes
ea449505 724X<character class> X<\p> X<\p{}>
ea449505
KW
725X<alpha> X<alnum> X<ascii> X<blank> X<cntrl> X<digit> X<graph>
726X<lower> X<print> X<punct> X<space> X<upper> X<word> X<xdigit>
8a118206 727
d66e1f56 728POSIX character classes have the form C<[:class:]>, where I<class> is the
6b83a163 729name, and the C<[:> and C<:]> delimiters. POSIX character classes only appear
8a118206 730I<inside> bracketed character classes, and are a convenient and descriptive
82206b5e 731way of listing a group of characters.
6b83a163
KW
732
733Be careful about the syntax,
8a118206
RGS
734
735 # Correct:
736 $string =~ /[[:alpha:]]/
737
738 # Incorrect (will warn):
739 $string =~ /[:alpha:]/
740
741The latter pattern would be a character class consisting of a colon,
742and the letters C<a>, C<l>, C<p> and C<h>.
d66e1f56 743
82206b5e 744POSIX character classes can be part of a larger bracketed character class.
b6538e4f 745For example,
ea449505
KW
746
747 [01[:alpha:]%]
748
749is valid and matches '0', '1', any alphabetic character, and the percent sign.
8a118206
RGS
750
751Perl recognizes the following POSIX character classes:
752
ea449505 753 alpha Any alphabetical character ("[A-Za-z]").
48cbae4f 754 alnum Any alphanumeric character ("[A-Za-z0-9]").
ea449505 755 ascii Any character in the ASCII character set.
ea8b8ad2 756 blank A GNU extension, equal to a space or a horizontal tab ("\t").
ea449505
KW
757 cntrl Any control character. See Note [2] below.
758 digit Any decimal digit ("[0-9]"), equivalent to "\d".
759 graph Any printable character, excluding a space. See Note [3] below.
760 lower Any lowercase character ("[a-z]").
761 print Any printable character, including a space. See Note [4] below.
c1c4ae3a 762 punct Any graphical character excluding "word" characters. Note [5].
d28d8023
KW
763 space Any whitespace character. "\s" including the vertical tab
764 ("\cK").
ea449505
KW
765 upper Any uppercase character ("[A-Z]").
766 word A Perl extension ("[A-Za-z0-9_]"), equivalent to "\w".
767 xdigit Any hexadecimal digit ("[0-9a-fA-F]").
768
93106464
KW
769Like the L<Unicode properties|/Unicode Properties>, most of the POSIX
770properties match the same regardless of whether case-insensitive (C</i>)
771matching is in effect or not. The two exceptions are C<[:upper:]> and
772C<[:lower:]>. Under C</i>, they each match the union of C<[:upper:]> and
773C<[:lower:]>.
774
ea449505
KW
775Most POSIX character classes have two Unicode-style C<\p> property
776counterparts. (They are not official Unicode properties, but Perl extensions
777derived from official Unicode properties.) The table below shows the relation
778between POSIX character classes and these counterparts.
779
780One counterpart, in the column labelled "ASCII-range Unicode" in
b6538e4f 781the table, matches only characters in the ASCII character set.
ea449505
KW
782
783The other counterpart, in the column labelled "Full-range Unicode", matches any
784appropriate characters in the full Unicode character set. For example,
b6538e4f 785C<\p{Alpha}> matches not just the ASCII alphabetic characters, but any
82206b5e 786character in the entire Unicode character set considered alphabetic.
582da942 787An entry in the column labelled "backslash sequence" is a (short)
5db9882c 788equivalent.
ea449505 789
cbc24f92
KW
790 [[:...:]] ASCII-range Full-range backslash Note
791 Unicode Unicode sequence
ea449505 792 -----------------------------------------------------
cbc24f92
KW
793 alpha \p{PosixAlpha} \p{XPosixAlpha}
794 alnum \p{PosixAlnum} \p{XPosixAlnum}
82206b5e 795 ascii \p{ASCII}
cbc24f92
KW
796 blank \p{PosixBlank} \p{XPosixBlank} \h [1]
797 or \p{HorizSpace} [1]
798 cntrl \p{PosixCntrl} \p{XPosixCntrl} [2]
799 digit \p{PosixDigit} \p{XPosixDigit} \d
800 graph \p{PosixGraph} \p{XPosixGraph} [3]
801 lower \p{PosixLower} \p{XPosixLower}
802 print \p{PosixPrint} \p{XPosixPrint} [4]
803 punct \p{PosixPunct} \p{XPosixPunct} [5]
804 \p{PerlSpace} \p{XPerlSpace} \s [6]
805 space \p{PosixSpace} \p{XPosixSpace} [6]
806 upper \p{PosixUpper} \p{XPosixUpper}
807 word \p{PosixWord} \p{XPosixWord} \w
82206b5e 808 xdigit \p{PosixXDigit} \p{XPosixXDigit}
8a118206
RGS
809
810=over 4
811
ea449505
KW
812=item [1]
813
814C<\p{Blank}> and C<\p{HorizSpace}> are synonyms.
815
816=item [2]
8a118206 817
ea449505 818Control characters don't produce output as such, but instead usually control
b6538e4f 819the terminal somehow: for example, newline and backspace are control characters.
93106464
KW
820On ASCII platforms, in the ASCII range, characters whose code points are
821between 0 and 31 inclusive, plus 127 (C<DEL>) are control characters; on
822EBCDIC platforms, their counterparts are control characters.
8a118206 823
ea449505 824=item [3]
8a118206
RGS
825
826Any character that is I<graphical>, that is, visible. This class consists
b6538e4f 827of all alphanumeric characters and all punctuation characters.
8a118206 828
ea449505 829=item [4]
8a118206 830
b6538e4f
TC
831All printable characters, which is the set of all graphical characters
832plus those whitespace characters which are not also controls.
ea449505 833
b6dac59a 834=item [5]
ea449505 835
b6538e4f 836C<\p{PosixPunct}> and C<[[:punct:]]> in the ASCII range match all
ea449505
KW
837non-controls, non-alphanumeric, non-space characters:
838C<[-!"#$%&'()*+,./:;<=E<gt>?@[\\\]^_`{|}~]> (although if a locale is in effect,
839it could alter the behavior of C<[[:punct:]]>).
840
cbc24f92
KW
841The similarly named property, C<\p{Punct}>, matches a somewhat different
842set in the ASCII range, namely
0be9b861
KW
843C<[-!"#%&'()*,./:;?@[\\\]_{}]>. That is, it is missing the nine
844characters C<[$+E<lt>=E<gt>^`|~]>.
6c5a041f
KW
845This is because Unicode splits what POSIX considers to be punctuation into two
846categories, Punctuation and Symbols.
847
e2cfb18c 848C<\p{XPosixPunct}> and (under Unicode rules) C<[[:punct:]]>, match what
765fa144
KW
849C<\p{PosixPunct}> matches in the ASCII range, plus what C<\p{Punct}>
850matches. This is different than strictly matching according to
851C<\p{Punct}>. Another way to say it is that
82206b5e
KW
852if Unicode rules are in effect, C<[[:punct:]]> matches all characters
853that Unicode considers punctuation, plus all ASCII-range characters that
854Unicode considers symbols.
8a118206 855
ea449505 856=item [6]
8a118206 857
7fa2fdc0 858C<\p{XPerlSpace}> and C<\p{Space}> match identically starting with Perl
d28d8023 859v5.18. In earlier versions, these differ only in that in non-locale
779cf272 860matching, C<\p{XPerlSpace}> did not match the vertical tab, C<\cK>.
d28d8023 861Same for the two ASCII-only range forms.
8a118206
RGS
862
863=back
864
ab6199be 865There are various other synonyms that can be used besides the names
4cb26c52 866listed in the table. For example, C<\p{XPosixAlpha}> can be written as
ab6199be 867C<\p{Alpha}>. All are listed in
d66e1f56 868L<perluniprops/Properties accessible through \p{} and \P{}>.
ab6199be
KW
869
870Both the C<\p> counterparts always assume Unicode rules are in effect.
871On ASCII platforms, this means they assume that the code points from 128
872to 255 are Latin-1, and that means that using them under locale rules is
873unwise unless the locale is guaranteed to be Latin-1 or UTF-8. In contrast, the
874POSIX character classes are useful under locale rules. They are
875affected by the actual rules in effect, as follows:
876
877=over
878
879=item If the C</a> modifier, is in effect ...
880
881Each of the POSIX classes matches exactly the same as their ASCII-range
882counterparts.
883
884=item otherwise ...
885
886=over
887
888=item For code points above 255 ...
889
890The POSIX class matches the same as its Full-range counterpart.
891
892=item For code points below 256 ...
893
894=over
895
896=item if locale rules are in effect ...
897
a145a423
KW
898The POSIX class matches according to the locale, except:
899
900=over
901
902=item C<word>
903
904also includes the platform's native underscore character, no matter what
8129baca 905the locale is.
ab6199be 906
a145a423
KW
907=item C<ascii>
908
909on platforms that don't have the POSIX C<ascii> extension, this matches
910just the platform's native ASCII-range characters.
911
912=item C<blank>
913
914on platforms that don't have the POSIX C<blank> extension, this matches
915just the platform's native tab and space characters.
916
917=back
918
4b9734bf 919=item if Unicode rules are in effect ...
ab6199be
KW
920
921The POSIX class matches the same as the Full-range counterpart.
922
923=item otherwise ...
924
925The POSIX class matches the same as the ASCII range counterpart.
926
927=back
928
929=back
930
931=back
932
933Which rules apply are determined as described in
934L<perlre/Which character set modifier is in effect?>.
935
936It is proposed to change this behavior in a future release of Perl so that
937whether or not Unicode rules are in effect would not change the
4b9734bf 938behavior: Outside of locale, the POSIX classes
ab6199be
KW
939would behave like their ASCII-range counterparts. If you wish to
940comment on this proposal, send email to C<perl5-porters@perl.org>.
cbc24f92 941
1f59b283 942=head4 Negation of POSIX character classes
ea449505 943X<character class, negation>
8a118206
RGS
944
945A Perl extension to the POSIX character class is the ability to
946negate it. This is done by prefixing the class name with a caret (C<^>).
947Some examples:
948
ea449505
KW
949 POSIX ASCII-range Full-range backslash
950 Unicode Unicode sequence
951 -----------------------------------------------------
cbc24f92
KW
952 [[:^digit:]] \P{PosixDigit} \P{XPosixDigit} \D
953 [[:^space:]] \P{PosixSpace} \P{XPosixSpace}
954 \P{PerlSpace} \P{XPerlSpace} \S
955 [[:^word:]] \P{PerlWord} \P{XPosixWord} \W
956
765fa144 957The backslash sequence can mean either ASCII- or Full-range Unicode,
82206b5e 958depending on various factors as described in L<perlre/Which character set modifier is in effect?>.
8a118206
RGS
959
960=head4 [= =] and [. .]
961
b6538e4f 962Perl recognizes the POSIX character classes C<[=class=]> and
82206b5e 963C<[.class.]>, but does not (yet?) support them. Any attempt to use
b6538e4f 964either construct raises an exception.
8a118206
RGS
965
966=head4 Examples
967
968 /[[:digit:]]/ # Matches a character that is a digit.
969 /[01[:lower:]]/ # Matches a character that is either a
970 # lowercase letter, or '0' or '1'.
c1c4ae3a 971 /[[:digit:][:^xdigit:]]/ # Matches a character that can be anything
bc943be5
KW
972 # except the letters 'a' to 'f' and 'A' to
973 # 'F'. This is because the main character
974 # class is composed of two POSIX character
975 # classes that are ORed together, one that
976 # matches any digit, and the other that
977 # matches anything that isn't a hex digit.
978 # The OR adds the digits, leaving only the
979 # letters 'a' to 'f' and 'A' to 'F' excluded.
572224ce
KW
980
981=head3 Extended Bracketed Character Classes
982X<character class>
983X<set operations>
984
985This is a fancy bracketed character class that can be used for more
986readable and less error-prone classes, and to perform set operations,
987such as intersection. An example is
988
989 /(?[ \p{Thai} & \p{Digit} ])/
990
991This will match all the digit characters that are in the Thai script.
992
993This is an experimental feature available starting in 5.18, and is
994subject to change as we gain field experience with it. Any attempt to
995use it will raise a warning, unless disabled via
996
997 no warnings "experimental::regex_sets";
998
999Comments on this feature are welcome; send email to
1000C<perl5-porters@perl.org>.
1001
a60b7922
KW
1002The rules used by L<C<use re 'strict>|re/'strict' mode> apply to this
1003construct.
1004
572224ce
KW
1005We can extend the example above:
1006
1007 /(?[ ( \p{Thai} + \p{Lao} ) & \p{Digit} ])/
1008
1009This matches digits that are in either the Thai or Laotian scripts.
1010
1011Notice the white space in these examples. This construct always has
d66e1f56 1012the C<E<sol>x> modifier turned on within it.
572224ce
KW
1013
1014The available binary operators are:
1015
1016 & intersection
1017 + union
1018 | another name for '+', hence means union
1019 - subtraction (the result matches the set consisting of those
1020 code points matched by the first operand, excluding any that
1021 are also matched by the second operand)
1022 ^ symmetric difference (the union minus the intersection). This
1023 is like an exclusive or, in that the result is the set of code
1024 points that are matched by either, but not both, of the
1025 operands.
1026
1027There is one unary operator:
1028
1029 ! complement
1030
6798c95d
KW
1031All the binary operators left associate; C<"&"> is higher precedence
1032than the others, which all have equal precedence. The unary operator
1033right associates, and has highest precedence. Thus this follows the
1034normal Perl precedence rules for logical operators. Use parentheses to
1035override the default precedence and associativity.
572224ce
KW
1036
1037The main restriction is that everything is a metacharacter. Thus,
1038you cannot refer to single characters by doing something like this:
1039
1040 /(?[ a + b ])/ # Syntax error!
1041
1042The easiest way to specify an individual typable character is to enclose
1043it in brackets:
1044
1045 /(?[ [a] + [b] ])/
1046
1047(This is the same thing as C<[ab]>.) You could also have said the
1048equivalent:
1049
1050 /(?[[ a b ]])/
1051
de36fb2e
KW
1052(You can, of course, specify single characters by using, C<\x{...}>,
1053C<\N{...}>, etc.)
572224ce
KW
1054
1055This last example shows the use of this construct to specify an ordinary
1056bracketed character class without additional set operations. Note the
d6b89212
KW
1057white space within it; a limited version of C<E<sol>x> is turned on even
1058within bracketed character classes, with only the SPACE and TAB (C<\t>)
1059characters allowed, and no comments. Hence,
572224ce
KW
1060
1061 (?[ [#] ])
1062
1063matches the literal character "#". To specify a literal white space character,
1064you can escape it with a backslash, like:
1065
1066 /(?[ [ a e i o u \ ] ])/
1067
1068This matches the English vowels plus the SPACE character.
1069All the other escapes accepted by normal bracketed character classes are
1070accepted here as well; but unrecognized escapes that generate warnings
1071in normal classes are fatal errors here.
1072
1073All warnings from these class elements are fatal, as well as some
1074practices that don't currently warn. For example you cannot say
1075
1076 /(?[ [ \xF ] ])/ # Syntax error!
1077
1078You have to have two hex digits after a braceless C<\x> (use a leading
1079zero to make two). These restrictions are to lower the incidence of
1080typos causing the class to not match what you thought it would.
1081
f194034a
KW
1082If a regular bracketed character class contains a C<\p{}> or C<\P{}> and
1083is matched against a non-Unicode code point, a warning may be
1084raised, as the result is not Unicode-defined. No such warning will come
1085when using this extended form.
1086
572224ce
KW
1087The final difference between regular bracketed character classes and
1088these, is that it is not possible to get these to match a
1089multi-character fold. Thus,
1090
1091 /(?[ [\xDF] ])/iu
1092
1093does not match the string C<ss>.
1094
1095You don't have to enclose POSIX class names inside double brackets,
1096hence both of the following work:
1097
1098 /(?[ [:word:] - [:lower:] ])/
1099 /(?[ [[:word:]] - [[:lower:]] ])/
1100
1101Any contained POSIX character classes, including things like C<\w> and C<\D>
1102respect the C<E<sol>a> (and C<E<sol>aa>) modifiers.
1103
1104C<< (?[ ]) >> is a regex-compile-time construct. Any attempt to use
1105something which isn't knowable at the time the containing regular
1106expression is compiled is a fatal error. In practice, this means
11a9b3e0 1107just three limitations:
572224ce
KW
1108
1109=over 4
1110
1111=item 1
1112
a0bd1a30
KW
1113When compiled within the scope of C<use locale> (or the C<E<sol>l> regex
1114modifier), this construct assumes that the execution-time locale will be
1115a UTF-8 one, and the generated pattern always uses Unicode rules. What
1116gets matched or not thus isn't dependent on the actual runtime locale, so
1117tainting is not enabled. But a C<locale> category warning is raised
1118if the runtime locale turns out to not be UTF-8.
572224ce
KW
1119
1120=item 2
1121
1122Any
1123L<user-defined property|perlunicode/"User-Defined Character Properties">
1124used must be already defined by the time the regular expression is
1125compiled (but note that this construct can be used instead of such
1126properties).
1127
1128=item 3
1129
1130A regular expression that otherwise would compile
1131using C<E<sol>d> rules, and which uses this construct will instead
1132use C<E<sol>u>. Thus this construct tells Perl that you don't want
1133C<E<sol>d> rules for the entire regular expression containing it.
1134
1135=back
1136
572224ce
KW
1137Note that skipping white space applies only to the interior of this
1138construct. There must not be any space between any of the characters
1139that form the initial C<(?[>. Nor may there be space between the
1140closing C<])> characters.
1141
11a9b3e0 1142Just as in all regular expressions, the pattern can be built up by
572224ce
KW
1143including variables that are interpolated at regex compilation time.
1144Care must be taken to ensure that you are getting what you expect. For
1145example:
1146
1147 my $thai_or_lao = '\p{Thai} + \p{Lao}';
1148 ...
1149 qr/(?[ \p{Digit} & $thai_or_lao ])/;
1150
1151compiles to
1152
1153 qr/(?[ \p{Digit} & \p{Thai} + \p{Lao} ])/;
1154
1155But this does not have the effect that someone reading the code would
1156likely expect, as the intersection applies just to C<\p{Thai}>,
1157excluding the Laotian. Pitfalls like this can be avoided by
1158parenthesizing the component pieces:
1159
1160 my $thai_or_lao = '( \p{Thai} + \p{Lao} )';
1161
1162But any modifiers will still apply to all the components:
1163
1164 my $lower = '\p{Lower} + \p{Digit}';
1165 qr/(?[ \p{Greek} & $lower ])/i;
1166
1167matches upper case things. You can avoid surprises by making the
1168components into instances of this construct by compiling them:
1169
1170 my $thai_or_lao = qr/(?[ \p{Thai} + \p{Lao} ])/;
1171 my $lower = qr/(?[ \p{Lower} + \p{Digit} ])/;
1172
1173When these are embedded in another pattern, what they match does not
1174change, regardless of parenthesization or what modifiers are in effect
1175in that outer pattern.
1176
1177Due to the way that Perl parses things, your parentheses and brackets
1178may need to be balanced, even including comments. If you run into any
1179examples, please send them to C<perlbug@perl.org>, so that we can have a
1180concrete example for this man page.
1181
1182We may change it so that things that remain legal uses in normal bracketed
1183character classes might become illegal within this experimental
1184construct. One proposal, for example, is to forbid adjacent uses of the
1185same character, as in C<(?[ [aa] ])>. The motivation for such a change
1186is that this usage is likely a typo, as the second "a" adds nothing.