This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Add Unicode property wildcards
[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
4a88d526 30add matching the newline by using the I<single line> modifier:
6b83a163 31for the entire regular expression with the C</s> modifier, or
4a88d526
KW
32locally with C<(?s)> (and even globally within the scope of
33L<C<use re '/s'>|re/'E<sol>flags' mode>). (The C<L</\N>> backslash
34sequence, described
6b83a163
KW
35below, matches any character except newline without regard to the
36I<single line> modifier.)
8a118206
RGS
37
38Here are some examples:
39
40 "a" =~ /./ # Match
41 "." =~ /./ # Match
42 "" =~ /./ # No match (dot has to match a character)
43 "\n" =~ /./ # No match (dot does not match a newline)
44 "\n" =~ /./s # Match (global 'single line' modifier)
45 "\n" =~ /(?s:.)/ # Match (local 'single line' modifier)
46 "ab" =~ /^.$/ # No match (dot matches one character)
47
6b83a163 48=head2 Backslash sequences
82206b5e 49X<\w> X<\W> X<\s> X<\S> X<\d> X<\D> X<\p> X<\P>
ea449505
KW
50X<\N> X<\v> X<\V> X<\h> X<\H>
51X<word> X<whitespace>
8a118206 52
6b83a163
KW
53A backslash sequence is a sequence of characters, the first one of which is a
54backslash. Perl ascribes special meaning to many such sequences, and some of
55these are character classes. That is, they match a single character each,
56provided that the character belongs to the specific set of characters defined
57by the sequence.
8a118206 58
6b83a163
KW
59Here's a list of the backslash sequences that are character classes. They
60are discussed in more detail below. (For the backslash sequences that aren't
61character classes, see L<perlrebackslash>.)
8a118206 62
6b83a163
KW
63 \d Match a decimal digit character.
64 \D Match a non-decimal-digit character.
8a118206
RGS
65 \w Match a "word" character.
66 \W Match a non-"word" character.
ea449505
KW
67 \s Match a whitespace character.
68 \S Match a non-whitespace character.
69 \h Match a horizontal whitespace character.
70 \H Match a character that isn't horizontal whitespace.
ea449505
KW
71 \v Match a vertical whitespace character.
72 \V Match a character that isn't vertical whitespace.
4e5e0888 73 \N Match a character that isn't a newline.
6b83a163 74 \pP, \p{Prop} Match a character that has the given Unicode property.
6c5a041f 75 \PP, \P{Prop} Match a character that doesn't have the Unicode property
8a118206 76
1433f837
KW
77=head3 \N
78
2171640d 79C<\N>, available starting in v5.12, like the dot, matches any
1433f837
KW
80character that is not a newline. The difference is that C<\N> is not influenced
81by the I<single line> regular expression modifier (see L</The dot> above). Note
82that the form C<\N{...}> may mean something completely different. When the
83C<{...}> is a L<quantifier|perlre/Quantifiers>, it means to match a non-newline
84character that many times. For example, C<\N{3}> means to match 3
85non-newlines; C<\N{5,}> means to match 5 or more non-newlines. But if C<{...}>
86is not a legal quantifier, it is presumed to be a named character. See
87L<charnames> for those. For example, none of C<\N{COLON}>, C<\N{4F}>, and
88C<\N{F4}> contain legal quantifiers, so Perl will try to find characters whose
89names are respectively C<COLON>, C<4F>, and C<F4>.
90
8a118206
RGS
91=head3 Digits
92
b6538e4f 93C<\d> matches a single character considered to be a decimal I<digit>.
5db9882c 94If the C</a> regular expression modifier is in effect, it matches [0-9].
582da942 95Otherwise, it
82206b5e
KW
96matches anything that is matched by C<\p{Digit}>, which includes [0-9].
97(An unlikely possible exception is that under locale matching rules, the
d66e1f56
KW
98current locale might not have C<[0-9]> matched by C<\d>, and/or might match
99other characters whose code point is less than 256. The only such locale
100definitions that are legal would be to match C<[0-9]> plus another set of
10110 consecutive digit characters; anything else would be in violation of
102the C language standard, but Perl doesn't currently assume anything in
103regard to this.)
82206b5e
KW
104
105What this means is that unless the C</a> modifier is in effect C<\d> not
106only matches the digits '0' - '9', but also Arabic, Devanagari, and
107digits from other languages. This may cause some confusion, and some
108security issues.
109
110Some digits that C<\d> matches look like some of the [0-9] ones, but
111have different values. For example, BENGALI DIGIT FOUR (U+09EA) looks
8350b274
KW
112very much like an ASCII DIGIT EIGHT (U+0038), and LEPCHA DIGIT SIX
113(U+1C46) looks very much like an ASCII DIGIT FIVE (U+0035). An
114application that
82206b5e
KW
115is expecting only the ASCII digits might be misled, or if the match is
116C<\d+>, the matched string might contain a mixture of digits from
117different writing systems that look like they signify a number different
67592e11 118than they actually do. L<Unicode::UCD/num()> can
e397bccf 119be used to safely
82206b5e 120calculate the value, returning C<undef> if the input string contains
8350b274
KW
121such a mixture. Otherwise, for example, a displayed price might be
122deliberately different than it appears.
82206b5e
KW
123
124What C<\p{Digit}> means (and hence C<\d> except under the C</a>
125modifier) is C<\p{General_Category=Decimal_Number}>, or synonymously,
126C<\p{General_Category=Digit}>. Starting with Unicode version 4.1, this
127is the same set of characters matched by C<\p{Numeric_Type=Decimal}>.
6b83a163
KW
128But Unicode also has a different property with a similar name,
129C<\p{Numeric_Type=Digit}>, which matches a completely different set of
82206b5e
KW
130characters. These characters are things such as C<CIRCLED DIGIT ONE>
131or subscripts, or are from writing systems that lack all ten digits.
6b83a163 132
82206b5e
KW
133The design intent is for C<\d> to exactly match the set of characters
134that can safely be used with "normal" big-endian positional decimal
135syntax, where, for example 123 means one 'hundred', plus two 'tens',
136plus three 'ones'. This positional notation does not necessarily apply
137to characters that match the other type of "digit",
138C<\p{Numeric_Type=Digit}>, and so C<\d> doesn't match them.
6b83a163 139
e2cfb18c 140The Tamil digits (U+0BE6 - U+0BEF) can also legally be
82206b5e
KW
141used in old-style Tamil numbers in which they would appear no more than
142one in a row, separated by characters that mean "times 10", "times 100",
143etc. (See L<http://www.unicode.org/notes/tn21>.)
8a118206 144
b6538e4f 145Any character not matched by C<\d> is matched by C<\D>.
8a118206
RGS
146
147=head3 Word characters
148
ea449505 149A C<\w> matches a single alphanumeric character (an alphabetic character, or a
41805eb9
KW
150decimal digit); or a connecting punctuation character, such as an
151underscore ("_"); or a "mark" character (like some sort of accent) that
152attaches to one of those. It does not match a whole word. To match a
153whole word, use C<\w+>. This isn't the same thing as matching an
154English word, but in the ASCII range it is the same as a string of
155Perl-identifier characters.
82206b5e
KW
156
157=over
158
159=item If the C</a> modifier is in effect ...
160
161C<\w> matches the 63 characters [a-zA-Z0-9_].
162
163=item otherwise ...
164
165=over
166
167=item For code points above 255 ...
168
169C<\w> matches the same as C<\p{Word}> matches in this range. That is,
170it matches Thai letters, Greek letters, etc. This includes connector
d35dd6c6 171punctuation (like the underscore) which connect two words together, or
b6538e4f 172diacritics, such as a C<COMBINING TILDE> and the modifier letters, which
82206b5e
KW
173are generally used to add auxiliary markings to letters.
174
175=item For code points below 256 ...
176
177=over
178
179=item if locale rules are in effect ...
180
181C<\w> matches the platform's native underscore character plus whatever
182the locale considers to be alphanumeric.
183
04c2d19c 184=item if, instead, Unicode rules are in effect ...
82206b5e
KW
185
186C<\w> matches exactly what C<\p{Word}> matches.
187
188=item otherwise ...
189
190C<\w> matches [a-zA-Z0-9_].
191
192=back
193
194=back
195
196=back
197
198Which rules apply are determined as described in L<perlre/Which character set modifier is in effect?>.
8a118206 199
6b83a163
KW
200There are a number of security issues with the full Unicode list of word
201characters. See L<http://unicode.org/reports/tr36>.
202
203Also, for a somewhat finer-grained set of characters that are in programming
204language identifiers beyond the ASCII range, you may wish to instead use the
e2cfb18c
KW
205more customized L</Unicode Properties>, C<\p{ID_Start}>,
206C<\p{ID_Continue}>, C<\p{XID_Start}>, and C<\p{XID_Continue}>. See
207L<http://unicode.org/reports/tr31>.
6b83a163 208
b6538e4f 209Any character not matched by C<\w> is matched by C<\W>.
8a118206 210
ea449505
KW
211=head3 Whitespace
212
82206b5e
KW
213C<\s> matches any single character considered whitespace.
214
215=over
216
217=item If the C</a> modifier is in effect ...
218
d28d8023
KW
219In all Perl versions, C<\s> matches the 5 characters [\t\n\f\r ]; that
220is, the horizontal tab,
221the newline, the form feed, the carriage return, and the space.
779cf272 222Starting in Perl v5.18, it also matches the vertical tab, C<\cK>.
d28d8023 223See note C<[1]> below for a discussion of this.
82206b5e
KW
224
225=item otherwise ...
226
227=over
228
229=item For code points above 255 ...
230
231C<\s> matches exactly the code points above 255 shown with an "s" column
232in the table below.
233
234=item For code points below 256 ...
235
236=over
237
238=item if locale rules are in effect ...
239
d28d8023 240C<\s> matches whatever the locale considers to be whitespace.
82206b5e 241
04c2d19c 242=item if, instead, Unicode rules are in effect ...
82206b5e
KW
243
244C<\s> matches exactly the characters shown with an "s" column in the
245table below.
246
247=item otherwise ...
248
779cf272 249C<\s> matches [\t\n\f\r ] and, starting in Perl
d28d8023
KW
250v5.18, the vertical tab, C<\cK>.
251(See note C<[1]> below for a discussion of this.)
82206b5e
KW
252Note that this list doesn't include the non-breaking space.
253
254=back
255
256=back
257
258=back
259
260Which rules apply are determined as described in L<perlre/Which character set modifier is in effect?>.
8a118206 261
b6538e4f 262Any character not matched by C<\s> is matched by C<\S>.
8a118206 263
b6538e4f 264C<\h> matches any character considered horizontal whitespace;
8129baca 265this includes the platform's space and tab characters and several others
b6538e4f 266listed in the table below. C<\H> matches any character
8129baca
KW
267not considered horizontal whitespace. They use the platform's native
268character set, and do not consider any locale that may otherwise be in
269use.
ea449505 270
b6538e4f 271C<\v> matches any character considered vertical whitespace;
8129baca 272this includes the platform's carriage return and line feed characters (newline)
b6538e4f
TC
273plus several other characters, all listed in the table below.
274C<\V> matches any character not considered vertical whitespace.
8129baca
KW
275They use the platform's native character set, and do not consider any
276locale that may otherwise be in use.
8a118206
RGS
277
278C<\R> matches anything that can be considered a newline under Unicode
412a49a2
KW
279rules. It can match a multi-character sequence. It cannot be used inside
280a bracketed character class; use C<\v> instead (vertical whitespace).
281It uses the platform's
8129baca
KW
282native character set, and does not consider any locale that may
283otherwise be in use.
ea449505 284Details are discussed in L<perlrebackslash>.
8a118206 285
82206b5e 286Note that unlike C<\s> (and C<\d> and C<\w>), C<\h> and C<\v> always match
8129baca
KW
287the same characters, without regard to other factors, such as the active
288locale or whether the source string is in UTF-8 format.
8a118206 289
d28d8023
KW
290One might think that C<\s> is equivalent to C<[\h\v]>. This is indeed true
291starting in Perl v5.18, but prior to that, the sole difference was that the
292vertical tab (C<"\cK">) was not matched by C<\s>.
8a118206
RGS
293
294The following table is a complete listing of characters matched by
a9c9e371 295C<\s>, C<\h> and C<\v> as of Unicode 6.3.
8a118206 296
582da942 297The first column gives the Unicode code point of the character (in hex format),
8a118206 298the second column gives the (Unicode) name. The third column indicates
4b9734bf
KW
299by which class(es) the character is matched (assuming no locale is in
300effect that changes the C<\s> matching).
8a118206 301
fc28d2a3
KW
302 0x0009 CHARACTER TABULATION h s
303 0x000a LINE FEED (LF) vs
d28d8023 304 0x000b LINE TABULATION vs [1]
fc28d2a3
KW
305 0x000c FORM FEED (FF) vs
306 0x000d CARRIAGE RETURN (CR) vs
307 0x0020 SPACE h s
d28d8023
KW
308 0x0085 NEXT LINE (NEL) vs [2]
309 0x00a0 NO-BREAK SPACE h s [2]
fc28d2a3 310 0x1680 OGHAM SPACE MARK h s
fc28d2a3
KW
311 0x2000 EN QUAD h s
312 0x2001 EM QUAD h s
313 0x2002 EN SPACE h s
314 0x2003 EM SPACE h s
315 0x2004 THREE-PER-EM SPACE h s
316 0x2005 FOUR-PER-EM SPACE h s
317 0x2006 SIX-PER-EM SPACE h s
318 0x2007 FIGURE SPACE h s
319 0x2008 PUNCTUATION SPACE h s
320 0x2009 THIN SPACE h s
321 0x200a HAIR SPACE h s
322 0x2028 LINE SEPARATOR vs
323 0x2029 PARAGRAPH SEPARATOR vs
324 0x202f NARROW NO-BREAK SPACE h s
325 0x205f MEDIUM MATHEMATICAL SPACE h s
326 0x3000 IDEOGRAPHIC SPACE h s
8a118206
RGS
327
328=over 4
329
330=item [1]
331
779cf272
KW
332Prior to Perl v5.18, C<\s> did not match the vertical tab.
333C<[^\S\cK]> (obscurely) matches what C<\s> traditionally did.
d28d8023
KW
334
335=item [2]
336
82206b5e
KW
337NEXT LINE and NO-BREAK SPACE may or may not match C<\s> depending
338on the rules in effect. See
339L<the beginning of this section|/Whitespace>.
8a118206
RGS
340
341=back
342
8a118206
RGS
343=head3 Unicode Properties
344
c1c4ae3a
KW
345C<\pP> and C<\p{Prop}> are character classes to match characters that fit given
346Unicode properties. One letter property names can be used in the C<\pP> form,
347with the property name following the C<\p>, otherwise, braces are required.
348When using braces, there is a single form, which is just the property name
349enclosed in the braces, and a compound form which looks like C<\p{name=value}>,
b6538e4f 350which means to match if the property "name" for the character has that particular
c1c4ae3a 351"value".
e1b711da
KW
352For instance, a match for a number can be written as C</\pN/> or as
353C</\p{Number}/>, or as C</\p{Number=True}/>.
354Lowercase letters are matched by the property I<Lowercase_Letter> which
e2cfb18c 355has the short form I<Ll>. They need the braces, so are written as C</\p{Ll}/> or
e1b711da
KW
356C</\p{Lowercase_Letter}/>, or C</\p{General_Category=Lowercase_Letter}/>
357(the underscores are optional).
358C</\pLl/> is valid, but means something different.
8a118206
RGS
359It matches a two character string: a letter (Unicode property C<\pL>),
360followed by a lowercase C<l>.
361
bc943be5 362If locale rules are not in effect, the use of
82206b5e 363a Unicode property will force the regular expression into using Unicode
bc943be5 364rules, if it isn't already.
82206b5e 365
56ca34ca
KW
366Note that almost all properties are immune to case-insensitive matching.
367That is, adding a C</i> regular expression modifier does not change what
82206b5e 368they match. There are two sets that are affected. The first set is
56ca34ca
KW
369C<Uppercase_Letter>,
370C<Lowercase_Letter>,
371and C<Titlecase_Letter>,
372all of which match C<Cased_Letter> under C</i> matching.
b6538e4f 373The second set is
56ca34ca
KW
374C<Uppercase>,
375C<Lowercase>,
376and C<Titlecase>,
377all of which match C<Cased> under C</i> matching.
378(The difference between these sets is that some things, such as Roman
e2cfb18c 379numerals, come in both upper and lower case, so they are C<Cased>, but
b6538e4f 380aren't considered to be letters, so they aren't C<Cased_Letter>s. They're
82206b5e
KW
381actually C<Letter_Number>s.)
382This set also includes its subsets C<PosixUpper> and C<PosixLower>, both
e2cfb18c 383of which under C</i> match C<PosixAlpha>.
56ca34ca
KW
384
385For more details on Unicode properties, see L<perlunicode/Unicode
386Character Properties>; for a
e1b711da 387complete list of possible properties, see
56ca34ca
KW
388L<perluniprops/Properties accessible through \p{} and \P{}>,
389which notes all forms that have C</i> differences.
e1b711da 390It is also possible to define your own properties. This is discussed in
8a118206
RGS
391L<perlunicode/User-Defined Character Properties>.
392
94b42e47 393Unicode properties are defined (surprise!) only on Unicode code points.
2d88a86a
KW
394Starting in v5.20, when matching against C<\p> and C<\P>, Perl treats
395non-Unicode code points (those above the legal Unicode maximum of
3960x10FFFF) as if they were typical unassigned Unicode code points.
94b42e47 397
2d88a86a
KW
398Prior to v5.20, Perl raised a warning and made all matches fail on
399non-Unicode code points. This could be somewhat surprising:
94b42e47 400
2d88a86a
KW
401 chr(0x110000) =~ \p{ASCII_Hex_Digit=True} # Fails on Perls < v5.20.
402 chr(0x110000) =~ \p{ASCII_Hex_Digit=False} # Also fails on Perls
403 # < v5.20
404
405Even though these two matches might be thought of as complements, until
406v5.20 they were so only on Unicode code points.
94b42e47 407
1532347b
KW
408Starting in perl v5.30, wildcards are allowed in Unicode property
409values. See L<perlunicode/Wildcards in Property Values>.
410
8a118206
RGS
411=head4 Examples
412
413 "a" =~ /\w/ # Match, "a" is a 'word' character.
414 "7" =~ /\w/ # Match, "7" is a 'word' character as well.
415 "a" =~ /\d/ # No match, "a" isn't a digit.
416 "7" =~ /\d/ # Match, "7" is a digit.
ea449505 417 " " =~ /\s/ # Match, a space is whitespace.
8a118206
RGS
418 "a" =~ /\D/ # Match, "a" is a non-digit.
419 "7" =~ /\D/ # No match, "7" is not a non-digit.
ea449505 420 " " =~ /\S/ # No match, a space is not non-whitespace.
8a118206 421
ea449505
KW
422 " " =~ /\h/ # Match, space is horizontal whitespace.
423 " " =~ /\v/ # No match, space is not vertical whitespace.
424 "\r" =~ /\v/ # Match, a return is vertical whitespace.
8a118206
RGS
425
426 "a" =~ /\pL/ # Match, "a" is a letter.
427 "a" =~ /\p{Lu}/ # No match, /\p{Lu}/ matches upper case letters.
428
429 "\x{0e0b}" =~ /\p{Thai}/ # Match, \x{0e0b} is the character
430 # 'THAI CHARACTER SO SO', and that's in
431 # Thai Unicode class.
ea449505 432 "a" =~ /\P{Lao}/ # Match, as "a" is not a Laotian character.
8a118206 433
82206b5e
KW
434It is worth emphasizing that C<\d>, C<\w>, etc, match single characters, not
435complete numbers or words. To match a number (that consists of digits),
436use C<\d+>; to match a word, use C<\w+>. But be aware of the security
437considerations in doing so, as mentioned above.
8a118206
RGS
438
439=head2 Bracketed Character Classes
440
441The third form of character class you can use in Perl regular expressions
6b83a163 442is the bracketed character class. In its simplest form, it lists the characters
c1c4ae3a 443that may be matched, surrounded by square brackets, like this: C<[aeiou]>.
ea449505 444This matches one of C<a>, C<e>, C<i>, C<o> or C<u>. Like the other
1f59b283 445character classes, exactly one character is matched.* To match
ea449505 446a longer string consisting of characters mentioned in the character
6b83a163 447class, follow the character class with a L<quantifier|perlre/Quantifiers>. For
b6538e4f 448instance, C<[aeiou]+> matches one or more lowercase English vowels.
8a118206
RGS
449
450Repeating a character in a character class has no
451effect; it's considered to be in the set only once.
452
453Examples:
454
455 "e" =~ /[aeiou]/ # Match, as "e" is listed in the class.
456 "p" =~ /[aeiou]/ # No match, "p" is not listed in the class.
457 "ae" =~ /^[aeiou]$/ # No match, a character class only matches
458 # a single character.
459 "ae" =~ /^[aeiou]+$/ # Match, due to the quantifier.
460
1f59b283
KW
461 -------
462
8f0cd35a
KW
463* There are two exceptions to a bracketed character class matching a
464single character only. Each requires special handling by Perl to make
465things work:
466
467=over
468
469=item *
470
471When the class is to match caselessly under C</i> matching rules, and a
472character that is explicitly mentioned inside the class matches a
1f59b283 473multiple-character sequence caselessly under Unicode rules, the class
8f0cd35a
KW
474will also match that sequence. For example, Unicode says that the
475letter C<LATIN SMALL LETTER SHARP S> should match the sequence C<ss>
476under C</i> rules. Thus,
1f59b283
KW
477
478 'ss' =~ /\A\N{LATIN SMALL LETTER SHARP S}\z/i # Matches
479 'ss' =~ /\A[aeioust\N{LATIN SMALL LETTER SHARP S}]\z/i # Matches
480
8f0cd35a
KW
481For this to happen, the class must not be inverted (see L</Negation>)
482and the character must be explicitly specified, and not be part of a
483multi-character range (not even as one of its endpoints). (L</Character
484Ranges> will be explained shortly.) Therefore,
9d53c457 485
eb9e3b14
KW
486 'ss' =~ /\A[\0-\x{ff}]\z/ui # Doesn't match
487 'ss' =~ /\A[\0-\N{LATIN SMALL LETTER SHARP S}]\z/ui # No match
488 'ss' =~ /\A[\xDF-\xDF]\z/ui # Matches on ASCII platforms, since
a845303d 489 # \xDF is LATIN SMALL LETTER SHARP S,
8f0cd35a
KW
490 # and the range is just a single
491 # element
9d53c457
KW
492
493Note that it isn't a good idea to specify these types of ranges anyway.
494
8f0cd35a
KW
495=item *
496
497Some names known to C<\N{...}> refer to a sequence of multiple characters,
498instead of the usual single character. When one of these is included in
499the class, the entire sequence is matched. For example,
500
501 "\N{TAMIL LETTER KA}\N{TAMIL VOWEL SIGN AU}"
502 =~ / ^ [\N{TAMIL SYLLABLE KAU}] $ /x;
503
504matches, because C<\N{TAMIL SYLLABLE KAU}> is a named sequence
505consisting of the two characters matched against. Like the other
eb9e3b14 506instance where a bracketed class can match multiple characters, and for
8f0cd35a
KW
507similar reasons, the class must not be inverted, and the named sequence
508may not appear in a range, even one where it is both endpoints. If
4a88d526
KW
509these happen, it is a fatal error if the character class is within the
510scope of L<C<use re 'strict>|re/'strict' mode>, or within an extended
511L<C<(?[...])>|/Extended Bracketed Character Classes> class; otherwise
512only the first code point is used (with a C<regexp>-type warning
513raised).
8f0cd35a
KW
514
515=back
516
8a118206
RGS
517=head3 Special Characters Inside a Bracketed Character Class
518
519Most characters that are meta characters in regular expressions (that
df225385 520is, characters that carry a special meaning like C<.>, C<*>, or C<(>) lose
8a118206
RGS
521their special meaning and can be used inside a character class without
522the need to escape them. For instance, C<[()]> matches either an opening
523parenthesis, or a closing parenthesis, and the parens inside the character
6e16fd37
KW
524class don't group or capture. Be aware that, unless the pattern is
525evaluated in single-quotish context, variable interpolation will take
526place before the bracketed class is parsed:
527
528 $, = "\t| ";
529 $a =~ m'[$,]'; # single-quotish: matches '$' or ','
530 $a =~ q{[$,]}' # same
531 $a =~ m/[$,]/; # double-quotish: matches "\t", "|", or " "
8a118206
RGS
532
533Characters that may carry a special meaning inside a character class are:
534C<\>, C<^>, C<->, C<[> and C<]>, and are discussed below. They can be
535escaped with a backslash, although this is sometimes not needed, in which
536case the backslash may be omitted.
537
538The sequence C<\b> is special inside a bracketed character class. While
6b83a163 539outside the character class, C<\b> is an assertion indicating a point
8a118206
RGS
540that does not have either two word characters or two non-word characters
541on either side, inside a bracketed character class, C<\b> matches a
542backspace character.
543
df225385
KW
544The sequences
545C<\a>,
546C<\c>,
547C<\e>,
548C<\f>,
549C<\n>,
e526e8bb 550C<\N{I<NAME>}>,
765fa144 551C<\N{U+I<hex char>}>,
df225385
KW
552C<\r>,
553C<\t>,
554and
555C<\x>
06ee63cd 556are also special and have the same meanings as they do outside a
eb9e3b14 557bracketed character class.
df225385 558
ea449505
KW
559Also, a backslash followed by two or three octal digits is considered an octal
560number.
df225385 561
6b83a163
KW
562A C<[> is not special inside a character class, unless it's the start of a
563POSIX character class (see L</POSIX Character Classes> below). It normally does
564not need escaping.
8a118206 565
6b83a163
KW
566A C<]> is normally either the end of a POSIX character class (see
567L</POSIX Character Classes> below), or it signals the end of the bracketed
568character class. If you want to include a C<]> in the set of characters, you
569must generally escape it.
b6538e4f 570
8a118206
RGS
571However, if the C<]> is the I<first> (or the second if the first
572character is a caret) character of a bracketed character class, it
573does not denote the end of the class (as you cannot have an empty class)
574and is considered part of the set of characters that can be matched without
575escaping.
576
577Examples:
578
579 "+" =~ /[+?*]/ # Match, "+" in a character class is not special.
090752cc 580 "\cH" =~ /[\b]/ # Match, \b inside in a character class
c1c4ae3a 581 # is equivalent to a backspace.
090752cc 582 "]" =~ /[][]/ # Match, as the character class contains
8a118206
RGS
583 # both [ and ].
584 "[]" =~ /[[]]/ # Match, the pattern contains a character class
52f4d632 585 # containing just [, and the character class is
8a118206
RGS
586 # followed by a ].
587
77c8f263
KW
588=head3 Bracketed Character Classes and the C</xx> pattern modifier
589
590Normally SPACE and TAB characters have no special meaning inside a
591bracketed character class; they are just added to the list of characters
592matched by the class. But if the L<C</xx>|perlre/E<sol>x and E<sol>xx>
593pattern modifier is in effect, they are generally ignored and can be
594added to improve readability. They can't be added in the middle of a
595single construct:
596
597 / [ \x{10 FFFF} ] /xx # WRONG!
598
599The SPACE in the middle of the hex constant is illegal.
600
601To specify a literal SPACE character, you can escape it with a
602backslash, like:
603
604 /[ a e i o u \ ]/xx
605
606This matches the English vowels plus the SPACE character.
607
608For clarity, you should already have been using C<\t> to specify a
609literal tab, and C<\t> is unaffected by C</xx>.
610
8a118206
RGS
611=head3 Character Ranges
612
613It is not uncommon to want to match a range of characters. Luckily, instead
b6538e4f 614of listing all characters in the range, one may use the hyphen (C<->).
8a118206 615If inside a bracketed character class you have two characters separated
b6538e4f 616by a hyphen, it's treated as if all characters between the two were in
8a118206 617the class. For instance, C<[0-9]> matches any ASCII digit, and C<[a-m]>
e2cfb18c 618matches any lowercase letter from the first half of the ASCII alphabet.
8a118206
RGS
619
620Note that the two characters on either side of the hyphen are not
765fa144 621necessarily both letters or both digits. Any character is possible,
8a118206 622although not advisable. C<['-?]> contains a range of characters, but
b6538e4f 623most people will not know which characters that means. Furthermore,
8a118206
RGS
624such ranges may lead to portability problems if the code has to run on
625a platform that uses a different character set, such as EBCDIC.
626
ea449505
KW
627If a hyphen in a character class cannot syntactically be part of a range, for
628instance because it is the first or the last character of the character class,
b6538e4f
TC
629or if it immediately follows a range, the hyphen isn't special, and so is
630considered a character to be matched literally. If you want a hyphen in
631your set of characters to be matched and its position in the class is such
632that it could be considered part of a range, you must escape that hyphen
633with a backslash.
8a118206
RGS
634
635Examples:
636
637 [a-z] # Matches a character that is a lower case ASCII letter.
c1c4ae3a
KW
638 [a-fz] # Matches any letter between 'a' and 'f' (inclusive) or
639 # the letter 'z'.
8a118206
RGS
640 [-z] # Matches either a hyphen ('-') or the letter 'z'.
641 [a-f-m] # Matches any letter between 'a' and 'f' (inclusive), the
642 # hyphen ('-'), or the letter 'm'.
643 ['-?] # Matches any of the characters '()*+,-./0123456789:;<=>?
644 # (But not on an EBCDIC platform).
c7d25594
KW
645 [\N{APOSTROPHE}-\N{QUESTION MARK}]
646 # Matches any of the characters '()*+,-./0123456789:;<=>?
647 # even on an EBCDIC platform.
ad63362f 648 [\N{U+27}-\N{U+3F}] # Same. (U+27 is "'", and U+3F is "?")
c7d25594 649
dabde021 650As the final two examples above show, you can achieve portability to
c7d25594
KW
651non-ASCII platforms by using the C<\N{...}> form for the range
652endpoints. These indicate that the specified range is to be interpreted
653using Unicode values, so C<[\N{U+27}-\N{U+3F}]> means to match
654C<\N{U+27}>, C<\N{U+28}>, C<\N{U+29}>, ..., C<\N{U+3D}>, C<\N{U+3E}>,
655and C<\N{U+3F}>, whatever the native code point versions for those are.
b927b7e9
KW
656These are called "Unicode" ranges. If either end is of the C<\N{...}>
657form, the range is considered Unicode. A C<regexp> warning is raised
658under C<S<"use re 'strict'">> if the other endpoint is specified
659non-portably:
660
661 [\N{U+00}-\x09] # Warning under re 'strict'; \x09 is non-portable
662 [\N{U+00}-\t] # No warning;
663
664Both of the above match the characters C<\N{U+00}> C<\N{U+01}>, ...
665C<\N{U+08}>, C<\N{U+09}>, but the C<\x09> looks like it could be a
666mistake so the warning is raised (under C<re 'strict'>) for it.
c7d25594
KW
667
668Perl also guarantees that the ranges C<A-Z>, C<a-z>, C<0-9>, and any
09e43397 669subranges of these match what an English-only speaker would expect them
c7d25594
KW
670to match on any platform. That is, C<[A-Z]> matches the 26 ASCII
671uppercase letters;
09e43397
KW
672C<[a-z]> matches the 26 lowercase letters; and C<[0-9]> matches the 10
673digits. Subranges, like C<[h-k]>, match correspondingly, in this case
674just the four letters C<"h">, C<"i">, C<"j">, and C<"k">. This is the
675natural behavior on ASCII platforms where the code points (ordinal
676values) for C<"h"> through C<"k"> are consecutive integers (0x68 through
6770x6B). But special handling to achieve this may be needed on platforms
678with a non-ASCII native character set. For example, on EBCDIC
679platforms, the code point for C<"h"> is 0x88, C<"i"> is 0x89, C<"j"> is
6800x91, and C<"k"> is 0x92. Perl specially treats C<[h-k]> to exclude the
681seven code points in the gap: 0x8A through 0x90. This special handling is
682only invoked when the range is a subrange of one of the ASCII uppercase,
683lowercase, and digit ranges, AND each end of the range is expressed
684either as a literal, like C<"A">, or as a named character (C<\N{...}>,
685including the C<\N{U+...> form).
686
687EBCDIC Examples:
688
689 [i-j] # Matches either "i" or "j"
690 [i-\N{LATIN SMALL LETTER J}] # Same
691 [i-\N{U+6A}] # Same
692 [\N{U+69}-\N{U+6A}] # Same
693 [\x{89}-\x{91}] # Matches 0x89 ("i"), 0x8A .. 0x90, 0x91 ("j")
694 [i-\x{91}] # Same
695 [\x{89}-j] # Same
696 [i-J] # Matches, 0x89 ("i") .. 0xC1 ("J"); special
697 # handling doesn't apply because range is mixed
698 # case
8a118206
RGS
699
700=head3 Negation
701
702It is also possible to instead list the characters you do not want to
703match. You can do so by using a caret (C<^>) as the first character in the
b6538e4f 704character class. For instance, C<[^a-z]> matches any character that is not a
e2cfb18c
KW
705lowercase ASCII letter, which therefore includes more than a million
706Unicode code points. The class is said to be "negated" or "inverted".
8a118206
RGS
707
708This syntax make the caret a special character inside a bracketed character
709class, but only if it is the first character of the class. So if you want
82206b5e 710the caret as one of the characters to match, either escape the caret or
e2cfb18c 711else don't list it first.
8a118206 712
1f59b283 713In inverted bracketed character classes, Perl ignores the Unicode rules
8f0cd35a
KW
714that normally say that named sequence, and certain characters should
715match a sequence of multiple characters use under caseless C</i>
716matching. Following those rules could lead to highly confusing
717situations:
1f59b283 718
582da942 719 "ss" =~ /^[^\xDF]+$/ui; # Matches!
1f59b283
KW
720
721This should match any sequences of characters that aren't C<\xDF> nor
722what C<\xDF> matches under C</i>. C<"s"> isn't C<\xDF>, but Unicode
723says that C<"ss"> is what C<\xDF> matches under C</i>. So which one
724"wins"? Do you fail the match because the string has C<ss> or accept it
582da942 725because it has an C<s> followed by another C<s>? Perl has chosen the
8f0cd35a 726latter. (See note in L</Bracketed Character Classes> above.)
1f59b283 727
8a118206
RGS
728Examples:
729
730 "e" =~ /[^aeiou]/ # No match, the 'e' is listed.
731 "x" =~ /[^aeiou]/ # Match, as 'x' isn't a lowercase vowel.
732 "^" =~ /[^^]/ # No match, matches anything that isn't a caret.
733 "^" =~ /[x^]/ # Match, caret is not special here.
734
735=head3 Backslash Sequences
736
ea449505 737You can put any backslash sequence character class (with the exception of
765fa144 738C<\N> and C<\R>) inside a bracketed character class, and it will act just
b6538e4f
TC
739as if you had put all characters matched by the backslash sequence inside the
740character class. For instance, C<[a-f\d]> matches any decimal digit, or any
6b83a163
KW
741of the lowercase letters between 'a' and 'f' inclusive.
742
743C<\N> within a bracketed character class must be of the forms C<\N{I<name>}>
765fa144 744or C<\N{U+I<hex char>}>, and NOT be the form that matches non-newlines,
6b83a163
KW
745for the same reason that a dot C<.> inside a bracketed character class loses
746its special meaning: it matches nearly anything, which generally isn't what you
747want to happen.
df225385 748
8a118206
RGS
749
750Examples:
751
752 /[\p{Thai}\d]/ # Matches a character that is either a Thai
753 # character, or a digit.
754 /[^\p{Arabic}()]/ # Matches a character that is neither an Arabic
755 # character, nor a parenthesis.
756
757Backslash sequence character classes cannot form one of the endpoints
6b83a163
KW
758of a range. Thus, you can't say:
759
760 /[\p{Thai}-\d]/ # Wrong!
8a118206 761
6b83a163 762=head3 POSIX Character Classes
ea449505 763X<character class> X<\p> X<\p{}>
ea449505
KW
764X<alpha> X<alnum> X<ascii> X<blank> X<cntrl> X<digit> X<graph>
765X<lower> X<print> X<punct> X<space> X<upper> X<word> X<xdigit>
8a118206 766
d66e1f56 767POSIX character classes have the form C<[:class:]>, where I<class> is the
6b83a163 768name, and the C<[:> and C<:]> delimiters. POSIX character classes only appear
8a118206 769I<inside> bracketed character classes, and are a convenient and descriptive
82206b5e 770way of listing a group of characters.
6b83a163
KW
771
772Be careful about the syntax,
8a118206
RGS
773
774 # Correct:
775 $string =~ /[[:alpha:]]/
776
777 # Incorrect (will warn):
778 $string =~ /[:alpha:]/
779
780The latter pattern would be a character class consisting of a colon,
781and the letters C<a>, C<l>, C<p> and C<h>.
d66e1f56 782
82206b5e 783POSIX character classes can be part of a larger bracketed character class.
b6538e4f 784For example,
ea449505
KW
785
786 [01[:alpha:]%]
787
788is valid and matches '0', '1', any alphabetic character, and the percent sign.
8a118206
RGS
789
790Perl recognizes the following POSIX character classes:
791
8a0ab3a4
KW
792 alpha Any alphabetical character (e.g., [A-Za-z]).
793 alnum Any alphanumeric character (e.g., [A-Za-z0-9]).
ea449505 794 ascii Any character in the ASCII character set.
ea8b8ad2 795 blank A GNU extension, equal to a space or a horizontal tab ("\t").
ea449505 796 cntrl Any control character. See Note [2] below.
8a0ab3a4 797 digit Any decimal digit (e.g., [0-9]), equivalent to "\d".
ea449505 798 graph Any printable character, excluding a space. See Note [3] below.
8a0ab3a4 799 lower Any lowercase character (e.g., [a-z]).
ea449505 800 print Any printable character, including a space. See Note [4] below.
c1c4ae3a 801 punct Any graphical character excluding "word" characters. Note [5].
d28d8023
KW
802 space Any whitespace character. "\s" including the vertical tab
803 ("\cK").
8a0ab3a4
KW
804 upper Any uppercase character (e.g., [A-Z]).
805 word A Perl extension (e.g., [A-Za-z0-9_]), equivalent to "\w".
7835a09a 806 xdigit Any hexadecimal digit (e.g., [0-9a-fA-F]). Note [7].
ea449505 807
93106464
KW
808Like the L<Unicode properties|/Unicode Properties>, most of the POSIX
809properties match the same regardless of whether case-insensitive (C</i>)
810matching is in effect or not. The two exceptions are C<[:upper:]> and
811C<[:lower:]>. Under C</i>, they each match the union of C<[:upper:]> and
812C<[:lower:]>.
813
ea449505
KW
814Most POSIX character classes have two Unicode-style C<\p> property
815counterparts. (They are not official Unicode properties, but Perl extensions
816derived from official Unicode properties.) The table below shows the relation
817between POSIX character classes and these counterparts.
818
819One counterpart, in the column labelled "ASCII-range Unicode" in
b6538e4f 820the table, matches only characters in the ASCII character set.
ea449505
KW
821
822The other counterpart, in the column labelled "Full-range Unicode", matches any
823appropriate characters in the full Unicode character set. For example,
b6538e4f 824C<\p{Alpha}> matches not just the ASCII alphabetic characters, but any
82206b5e 825character in the entire Unicode character set considered alphabetic.
582da942 826An entry in the column labelled "backslash sequence" is a (short)
5db9882c 827equivalent.
ea449505 828
cbc24f92
KW
829 [[:...:]] ASCII-range Full-range backslash Note
830 Unicode Unicode sequence
ea449505 831 -----------------------------------------------------
cbc24f92
KW
832 alpha \p{PosixAlpha} \p{XPosixAlpha}
833 alnum \p{PosixAlnum} \p{XPosixAlnum}
82206b5e 834 ascii \p{ASCII}
cbc24f92
KW
835 blank \p{PosixBlank} \p{XPosixBlank} \h [1]
836 or \p{HorizSpace} [1]
837 cntrl \p{PosixCntrl} \p{XPosixCntrl} [2]
838 digit \p{PosixDigit} \p{XPosixDigit} \d
839 graph \p{PosixGraph} \p{XPosixGraph} [3]
840 lower \p{PosixLower} \p{XPosixLower}
841 print \p{PosixPrint} \p{XPosixPrint} [4]
842 punct \p{PosixPunct} \p{XPosixPunct} [5]
843 \p{PerlSpace} \p{XPerlSpace} \s [6]
844 space \p{PosixSpace} \p{XPosixSpace} [6]
845 upper \p{PosixUpper} \p{XPosixUpper}
846 word \p{PosixWord} \p{XPosixWord} \w
7835a09a 847 xdigit \p{PosixXDigit} \p{XPosixXDigit} [7]
8a118206
RGS
848
849=over 4
850
ea449505
KW
851=item [1]
852
853C<\p{Blank}> and C<\p{HorizSpace}> are synonyms.
854
855=item [2]
8a118206 856
ea449505 857Control characters don't produce output as such, but instead usually control
b6538e4f 858the terminal somehow: for example, newline and backspace are control characters.
93106464
KW
859On ASCII platforms, in the ASCII range, characters whose code points are
860between 0 and 31 inclusive, plus 127 (C<DEL>) are control characters; on
861EBCDIC platforms, their counterparts are control characters.
8a118206 862
ea449505 863=item [3]
8a118206
RGS
864
865Any character that is I<graphical>, that is, visible. This class consists
b6538e4f 866of all alphanumeric characters and all punctuation characters.
8a118206 867
ea449505 868=item [4]
8a118206 869
b6538e4f
TC
870All printable characters, which is the set of all graphical characters
871plus those whitespace characters which are not also controls.
ea449505 872
b6dac59a 873=item [5]
ea449505 874
b6538e4f 875C<\p{PosixPunct}> and C<[[:punct:]]> in the ASCII range match all
ea449505
KW
876non-controls, non-alphanumeric, non-space characters:
877C<[-!"#$%&'()*+,./:;<=E<gt>?@[\\\]^_`{|}~]> (although if a locale is in effect,
878it could alter the behavior of C<[[:punct:]]>).
879
cbc24f92
KW
880The similarly named property, C<\p{Punct}>, matches a somewhat different
881set in the ASCII range, namely
0be9b861
KW
882C<[-!"#%&'()*,./:;?@[\\\]_{}]>. That is, it is missing the nine
883characters C<[$+E<lt>=E<gt>^`|~]>.
6c5a041f
KW
884This is because Unicode splits what POSIX considers to be punctuation into two
885categories, Punctuation and Symbols.
886
e2cfb18c 887C<\p{XPosixPunct}> and (under Unicode rules) C<[[:punct:]]>, match what
765fa144
KW
888C<\p{PosixPunct}> matches in the ASCII range, plus what C<\p{Punct}>
889matches. This is different than strictly matching according to
890C<\p{Punct}>. Another way to say it is that
82206b5e
KW
891if Unicode rules are in effect, C<[[:punct:]]> matches all characters
892that Unicode considers punctuation, plus all ASCII-range characters that
893Unicode considers symbols.
8a118206 894
ea449505 895=item [6]
8a118206 896
7fa2fdc0 897C<\p{XPerlSpace}> and C<\p{Space}> match identically starting with Perl
d28d8023 898v5.18. In earlier versions, these differ only in that in non-locale
779cf272 899matching, C<\p{XPerlSpace}> did not match the vertical tab, C<\cK>.
d28d8023 900Same for the two ASCII-only range forms.
8a118206 901
7835a09a
KW
902=item [7]
903
904Unlike C<[[:digit:]]> which matches digits in many writing systems, such
905as Thai and Devanagari, there are currently only two sets of hexadecimal
906digits, and it is unlikely that more will be added. This is because you
907not only need the ten digits, but also the six C<[A-F]> (and C<[a-f]>)
908to correspond. That means only the Latin script is suitable for these,
909and Unicode has only two sets of these, the familiar ASCII set, and the
910fullwidth forms starting at U+FF10 (FULLWIDTH DIGIT ZERO).
911
8a118206
RGS
912=back
913
ab6199be 914There are various other synonyms that can be used besides the names
4cb26c52 915listed in the table. For example, C<\p{XPosixAlpha}> can be written as
ab6199be 916C<\p{Alpha}>. All are listed in
d66e1f56 917L<perluniprops/Properties accessible through \p{} and \P{}>.
ab6199be
KW
918
919Both the C<\p> counterparts always assume Unicode rules are in effect.
920On ASCII platforms, this means they assume that the code points from 128
921to 255 are Latin-1, and that means that using them under locale rules is
922unwise unless the locale is guaranteed to be Latin-1 or UTF-8. In contrast, the
923POSIX character classes are useful under locale rules. They are
924affected by the actual rules in effect, as follows:
925
926=over
927
928=item If the C</a> modifier, is in effect ...
929
930Each of the POSIX classes matches exactly the same as their ASCII-range
931counterparts.
932
933=item otherwise ...
934
935=over
936
937=item For code points above 255 ...
938
939The POSIX class matches the same as its Full-range counterpart.
940
941=item For code points below 256 ...
942
943=over
944
945=item if locale rules are in effect ...
946
a145a423
KW
947The POSIX class matches according to the locale, except:
948
949=over
950
951=item C<word>
952
953also includes the platform's native underscore character, no matter what
8129baca 954the locale is.
ab6199be 955
a145a423
KW
956=item C<ascii>
957
958on platforms that don't have the POSIX C<ascii> extension, this matches
959just the platform's native ASCII-range characters.
960
961=item C<blank>
962
963on platforms that don't have the POSIX C<blank> extension, this matches
964just the platform's native tab and space characters.
965
966=back
967
04c2d19c 968=item if, instead, Unicode rules are in effect ...
ab6199be
KW
969
970The POSIX class matches the same as the Full-range counterpart.
971
972=item otherwise ...
973
974The POSIX class matches the same as the ASCII range counterpart.
975
976=back
977
978=back
979
980=back
981
982Which rules apply are determined as described in
983L<perlre/Which character set modifier is in effect?>.
984
1f59b283 985=head4 Negation of POSIX character classes
ea449505 986X<character class, negation>
8a118206
RGS
987
988A Perl extension to the POSIX character class is the ability to
989negate it. This is done by prefixing the class name with a caret (C<^>).
990Some examples:
991
ea449505
KW
992 POSIX ASCII-range Full-range backslash
993 Unicode Unicode sequence
994 -----------------------------------------------------
cbc24f92
KW
995 [[:^digit:]] \P{PosixDigit} \P{XPosixDigit} \D
996 [[:^space:]] \P{PosixSpace} \P{XPosixSpace}
997 \P{PerlSpace} \P{XPerlSpace} \S
998 [[:^word:]] \P{PerlWord} \P{XPosixWord} \W
999
765fa144 1000The backslash sequence can mean either ASCII- or Full-range Unicode,
82206b5e 1001depending on various factors as described in L<perlre/Which character set modifier is in effect?>.
8a118206
RGS
1002
1003=head4 [= =] and [. .]
1004
b6538e4f 1005Perl recognizes the POSIX character classes C<[=class=]> and
82206b5e 1006C<[.class.]>, but does not (yet?) support them. Any attempt to use
b6538e4f 1007either construct raises an exception.
8a118206
RGS
1008
1009=head4 Examples
1010
1011 /[[:digit:]]/ # Matches a character that is a digit.
1012 /[01[:lower:]]/ # Matches a character that is either a
1013 # lowercase letter, or '0' or '1'.
c1c4ae3a 1014 /[[:digit:][:^xdigit:]]/ # Matches a character that can be anything
bc943be5
KW
1015 # except the letters 'a' to 'f' and 'A' to
1016 # 'F'. This is because the main character
1017 # class is composed of two POSIX character
1018 # classes that are ORed together, one that
1019 # matches any digit, and the other that
1020 # matches anything that isn't a hex digit.
1021 # The OR adds the digits, leaving only the
1022 # letters 'a' to 'f' and 'A' to 'F' excluded.
572224ce
KW
1023
1024=head3 Extended Bracketed Character Classes
1025X<character class>
1026X<set operations>
1027
1028This is a fancy bracketed character class that can be used for more
1029readable and less error-prone classes, and to perform set operations,
1030such as intersection. An example is
1031
1032 /(?[ \p{Thai} & \p{Digit} ])/
1033
1034This will match all the digit characters that are in the Thai script.
1035
1036This is an experimental feature available starting in 5.18, and is
1037subject to change as we gain field experience with it. Any attempt to
1038use it will raise a warning, unless disabled via
1039
1040 no warnings "experimental::regex_sets";
1041
1042Comments on this feature are welcome; send email to
1043C<perl5-porters@perl.org>.
1044
a60b7922
KW
1045The rules used by L<C<use re 'strict>|re/'strict' mode> apply to this
1046construct.
1047
572224ce
KW
1048We can extend the example above:
1049
1050 /(?[ ( \p{Thai} + \p{Lao} ) & \p{Digit} ])/
1051
1052This matches digits that are in either the Thai or Laotian scripts.
1053
1054Notice the white space in these examples. This construct always has
77c8f263 1055the C<E<sol>xx> modifier turned on within it.
572224ce
KW
1056
1057The available binary operators are:
1058
1059 & intersection
1060 + union
1061 | another name for '+', hence means union
1062 - subtraction (the result matches the set consisting of those
1063 code points matched by the first operand, excluding any that
1064 are also matched by the second operand)
1065 ^ symmetric difference (the union minus the intersection). This
1066 is like an exclusive or, in that the result is the set of code
1067 points that are matched by either, but not both, of the
1068 operands.
1069
1070There is one unary operator:
1071
1072 ! complement
1073
6798c95d
KW
1074All the binary operators left associate; C<"&"> is higher precedence
1075than the others, which all have equal precedence. The unary operator
1076right associates, and has highest precedence. Thus this follows the
1077normal Perl precedence rules for logical operators. Use parentheses to
1078override the default precedence and associativity.
572224ce
KW
1079
1080The main restriction is that everything is a metacharacter. Thus,
1081you cannot refer to single characters by doing something like this:
1082
1083 /(?[ a + b ])/ # Syntax error!
1084
1085The easiest way to specify an individual typable character is to enclose
1086it in brackets:
1087
1088 /(?[ [a] + [b] ])/
1089
1090(This is the same thing as C<[ab]>.) You could also have said the
1091equivalent:
1092
1093 /(?[[ a b ]])/
1094
de36fb2e
KW
1095(You can, of course, specify single characters by using, C<\x{...}>,
1096C<\N{...}>, etc.)
572224ce
KW
1097
1098This last example shows the use of this construct to specify an ordinary
1099bracketed character class without additional set operations. Note the
77c8f263
KW
1100white space within it. This is allowed because C<E<sol>xx> is
1101automatically turned on within this construct.
572224ce 1102
572224ce 1103All the other escapes accepted by normal bracketed character classes are
7d4c055d
KW
1104accepted here as well.
1105
1106Because this construct compiles under
1107L<C<use re 'strict>|re/'strict' mode>, unrecognized escapes that
1108generate warnings in normal classes are fatal errors here, as well as
1109all other warnings from these class elements, as well as some
1110practices that don't currently warn outside C<re 'strict'>. For example
1111you cannot say
572224ce
KW
1112
1113 /(?[ [ \xF ] ])/ # Syntax error!
1114
1115You have to have two hex digits after a braceless C<\x> (use a leading
1116zero to make two). These restrictions are to lower the incidence of
1117typos causing the class to not match what you thought it would.
1118
f194034a
KW
1119If a regular bracketed character class contains a C<\p{}> or C<\P{}> and
1120is matched against a non-Unicode code point, a warning may be
1121raised, as the result is not Unicode-defined. No such warning will come
1122when using this extended form.
1123
572224ce
KW
1124The final difference between regular bracketed character classes and
1125these, is that it is not possible to get these to match a
1126multi-character fold. Thus,
1127
1128 /(?[ [\xDF] ])/iu
1129
1130does not match the string C<ss>.
1131
1132You don't have to enclose POSIX class names inside double brackets,
1133hence both of the following work:
1134
1135 /(?[ [:word:] - [:lower:] ])/
1136 /(?[ [[:word:]] - [[:lower:]] ])/
1137
1138Any contained POSIX character classes, including things like C<\w> and C<\D>
1139respect the C<E<sol>a> (and C<E<sol>aa>) modifiers.
1140
19a498a4
YO
1141Note that C<< (?[ ]) >> is a regex-compile-time construct. Any attempt
1142to use something which isn't knowable at the time the containing regular
572224ce 1143expression is compiled is a fatal error. In practice, this means
11a9b3e0 1144just three limitations:
572224ce
KW
1145
1146=over 4
1147
1148=item 1
1149
a0bd1a30
KW
1150When compiled within the scope of C<use locale> (or the C<E<sol>l> regex
1151modifier), this construct assumes that the execution-time locale will be
1152a UTF-8 one, and the generated pattern always uses Unicode rules. What
1153gets matched or not thus isn't dependent on the actual runtime locale, so
1154tainting is not enabled. But a C<locale> category warning is raised
1155if the runtime locale turns out to not be UTF-8.
572224ce
KW
1156
1157=item 2
1158
1159Any
1160L<user-defined property|perlunicode/"User-Defined Character Properties">
1161used must be already defined by the time the regular expression is
1162compiled (but note that this construct can be used instead of such
1163properties).
1164
1165=item 3
1166
1167A regular expression that otherwise would compile
1168using C<E<sol>d> rules, and which uses this construct will instead
1169use C<E<sol>u>. Thus this construct tells Perl that you don't want
1170C<E<sol>d> rules for the entire regular expression containing it.
1171
1172=back
1173
572224ce
KW
1174Note that skipping white space applies only to the interior of this
1175construct. There must not be any space between any of the characters
1176that form the initial C<(?[>. Nor may there be space between the
1177closing C<])> characters.
1178
11a9b3e0 1179Just as in all regular expressions, the pattern can be built up by
572224ce 1180including variables that are interpolated at regex compilation time.
1c484ed4
KW
1181But its best to compile each sub-component.
1182
1183 my $thai_or_lao = qr/(?[ \p{Thai} + \p{Lao} ])/;
1184 my $lower = qr/(?[ \p{Lower} + \p{Digit} ])/;
1185
1186When these are embedded in another pattern, what they match does not
1187change, regardless of parenthesization or what modifiers are in effect
1188in that outer pattern. If you fail to compile the subcomponents, you
1189can get some nasty surprises. For example:
572224ce
KW
1190
1191 my $thai_or_lao = '\p{Thai} + \p{Lao}';
1192 ...
1193 qr/(?[ \p{Digit} & $thai_or_lao ])/;
1194
1195compiles to
1196
1197 qr/(?[ \p{Digit} & \p{Thai} + \p{Lao} ])/;
1198
1c484ed4
KW
1199But this does not have the effect that someone reading the source code
1200would likely expect, as the intersection applies just to C<\p{Thai}>,
1201excluding the Laotian. Its best to compile the subcomponents, but you
1202could also parenthesize the component pieces:
572224ce
KW
1203
1204 my $thai_or_lao = '( \p{Thai} + \p{Lao} )';
1205
1206But any modifiers will still apply to all the components:
1207
1208 my $lower = '\p{Lower} + \p{Digit}';
1209 qr/(?[ \p{Greek} & $lower ])/i;
1210
1c484ed4
KW
1211matches upper case things. So just, compile the subcomponents, as
1212illustrated above.
572224ce
KW
1213
1214Due to the way that Perl parses things, your parentheses and brackets
1215may need to be balanced, even including comments. If you run into any
1216examples, please send them to C<perlbug@perl.org>, so that we can have a
1217concrete example for this man page.
1218
1219We may change it so that things that remain legal uses in normal bracketed
1220character classes might become illegal within this experimental
1221construct. One proposal, for example, is to forbid adjacent uses of the
1222same character, as in C<(?[ [aa] ])>. The motivation for such a change
1223is that this usage is likely a typo, as the second "a" adds nothing.