character, except for the newline. That default can be changed to
add matching the newline by using the I<single line> modifier: either
for the entire regular expression with the C</s> modifier, or
-locally with C<(?s)>. (The experimental C<\N> backslash sequence, described
+locally with C<(?s)>. (The C<L</\N>> backslash sequence, described
below, matches any character except newline without regard to the
I<single line> modifier.)
\H Match a character that isn't horizontal whitespace.
\v Match a vertical whitespace character.
\V Match a character that isn't vertical whitespace.
- \N Match a character that isn't a newline. Experimental.
+ \N Match a character that isn't a newline.
\pP, \p{Prop} Match a character that has the given Unicode property.
\PP, \P{Prop} Match a character that doesn't have the Unicode property
=head3 \N
-C<\N> is new in 5.12, and is experimental. It, like the dot, matches any
+C<\N>, available starting in v5.12, like the dot, matches any
character that is not a newline. The difference is that C<\N> is not influenced
by the I<single line> regular expression modifier (see L</The dot> above). Note
that the form C<\N{...}> may mean something completely different. When the
Otherwise, it
matches anything that is matched by C<\p{Digit}>, which includes [0-9].
(An unlikely possible exception is that under locale matching rules, the
-current locale might not have [0-9] matched by C<\d>, and/or might match
-other characters whose code point is less than 256. Such a locale
-definition would be in violation of the C language standard, but Perl
-doesn't currently assume anything in regard to this.)
+current locale might not have C<[0-9]> matched by C<\d>, and/or might match
+other characters whose code point is less than 256. The only such locale
+definitions that are legal would be to match C<[0-9]> plus another set of
+10 consecutive digit characters; anything else would be in violation of
+the C language standard, but Perl doesn't currently assume anything in
+regard to this.)
What this means is that unless the C</a> modifier is in effect C<\d> not
only matches the digits '0' - '9', but also Arabic, Devanagari, and
=head3 Word characters
A C<\w> matches a single alphanumeric character (an alphabetic character, or a
-decimal digit) or a connecting punctuation character, such as an
-underscore ("_"). It does not match a whole word. To match a whole
-word, use C<\w+>. This isn't the same thing as matching an English word, but
-in the ASCII range it is the same as a string of Perl-identifier
-characters.
+decimal digit); or a connecting punctuation character, such as an
+underscore ("_"); or a "mark" character (like some sort of accent) that
+attaches to one of those. It does not match a whole word. To match a
+whole word, use C<\w+>. This isn't the same thing as matching an
+English word, but in the ASCII range it is the same as a string of
+Perl-identifier characters.
=over
C<\w> matches the platform's native underscore character plus whatever
the locale considers to be alphanumeric.
-=item if Unicode rules are in effect or if on an EBCDIC platform ...
+=item if Unicode rules are in effect ...
C<\w> matches exactly what C<\p{Word}> matches.
=item If the C</a> modifier is in effect ...
-C<\s> matches the 5 characters [\t\n\f\r ]; that is, the horizontal tab,
-the newline, the form feed, the carriage return, and the space. (Note
-that it doesn't match the vertical tab, C<\cK> on ASCII platforms.)
+In all Perl versions, C<\s> matches the 5 characters [\t\n\f\r ]; that
+is, the horizontal tab,
+the newline, the form feed, the carriage return, and the space.
+Starting in Perl v5.18, it also matches the vertical tab, C<\cK>.
+See note C<[1]> below for a discussion of this.
=item otherwise ...
=item if locale rules are in effect ...
-C<\s> matches whatever the locale considers to be whitespace. Note that
-this is likely to include the vertical space, unlike non-locale C<\s>
-matching.
+C<\s> matches whatever the locale considers to be whitespace.
-=item if Unicode rules are in effect or if on an EBCDIC platform ...
+=item if Unicode rules are in effect ...
C<\s> matches exactly the characters shown with an "s" column in the
table below.
=item otherwise ...
-C<\s> matches [\t\n\f\r ].
+C<\s> matches [\t\n\f\r ] and, starting in Perl
+v5.18, the vertical tab, C<\cK>.
+(See note C<[1]> below for a discussion of this.)
Note that this list doesn't include the non-breaking space.
=back
locale that may otherwise be in use.
C<\R> matches anything that can be considered a newline under Unicode
-rules. It's not a character class, as it can match a multi-character
-sequence. Therefore, it cannot be used inside a bracketed character
-class; use C<\v> instead (vertical whitespace). It uses the platform's
+rules. It can match a multi-character sequence. It cannot be used inside
+a bracketed character class; use C<\v> instead (vertical whitespace).
+It uses the platform's
native character set, and does not consider any locale that may
otherwise be in use.
Details are discussed in L<perlrebackslash>.
the same characters, without regard to other factors, such as the active
locale or whether the source string is in UTF-8 format.
-One might think that C<\s> is equivalent to C<[\h\v]>. This is not true.
-The difference is that the vertical tab (C<"\x0b">) is not matched by
-C<\s>; it is however considered vertical whitespace.
+One might think that C<\s> is equivalent to C<[\h\v]>. This is indeed true
+starting in Perl v5.18, but prior to that, the sole difference was that the
+vertical tab (C<"\cK">) was not matched by C<\s>.
The following table is a complete listing of characters matched by
-C<\s>, C<\h> and C<\v> as of Unicode 6.0.
+C<\s>, C<\h> and C<\v> as of Unicode 6.3.
The first column gives the Unicode code point of the character (in hex format),
the second column gives the (Unicode) name. The third column indicates
-by which class(es) the character is matched (assuming no locale or EBCDIC code
-page is in effect that changes the C<\s> matching).
+by which class(es) the character is matched (assuming no locale is in
+effect that changes the C<\s> matching).
0x0009 CHARACTER TABULATION h s
0x000a LINE FEED (LF) vs
- 0x000b LINE TABULATION v
+ 0x000b LINE TABULATION vs [1]
0x000c FORM FEED (FF) vs
0x000d CARRIAGE RETURN (CR) vs
0x0020 SPACE h s
- 0x0085 NEXT LINE (NEL) vs [1]
- 0x00a0 NO-BREAK SPACE h s [1]
+ 0x0085 NEXT LINE (NEL) vs [2]
+ 0x00a0 NO-BREAK SPACE h s [2]
0x1680 OGHAM SPACE MARK h s
- 0x180e MONGOLIAN VOWEL SEPARATOR h s
0x2000 EN QUAD h s
0x2001 EM QUAD h s
0x2002 EN SPACE h s
=item [1]
+Prior to Perl v5.18, C<\s> did not match the vertical tab.
+C<[^\S\cK]> (obscurely) matches what C<\s> traditionally did.
+
+=item [2]
+
NEXT LINE and NO-BREAK SPACE may or may not match C<\s> depending
on the rules in effect. See
L<the beginning of this section|/Whitespace>.
L<perlunicode/User-Defined Character Properties>.
Unicode properties are defined (surprise!) only on Unicode code points.
-A warning is raised and all matches fail on non-Unicode code points
-(those above the legal Unicode maximum of 0x10FFFF). This can be
-somewhat surprising,
+Starting in v5.20, when matching against C<\p> and C<\P>, Perl treats
+non-Unicode code points (those above the legal Unicode maximum of
+0x10FFFF) as if they were typical unassigned Unicode code points.
+
+Prior to v5.20, Perl raised a warning and made all matches fail on
+non-Unicode code points. This could be somewhat surprising:
- chr(0x110000) =~ \p{ASCII_Hex_Digit=True} # Fails.
- chr(0x110000) =~ \p{ASCII_Hex_Digit=False} # Also fails!
+ chr(0x110000) =~ \p{ASCII_Hex_Digit=True} # Fails on Perls < v5.20.
+ chr(0x110000) =~ \p{ASCII_Hex_Digit=False} # Also fails on Perls
+ # < v5.20
-Even though these two matches might be thought of as complements, they
-are so only on Unicode code points.
+Even though these two matches might be thought of as complements, until
+v5.20 they were so only on Unicode code points.
=head4 Examples
-------
-* There is an exception to a bracketed character class matching a
-single character only. When the class is to match caselessly under C</i>
-matching rules, and a character inside the class matches a
+* There are two exceptions to a bracketed character class matching a
+single character only. Each requires special handling by Perl to make
+things work:
+
+=over
+
+=item *
+
+When the class is to match caselessly under C</i> matching rules, and a
+character that is explicitly mentioned inside the class matches a
multiple-character sequence caselessly under Unicode rules, the class
-(when not L<inverted|/Negation>) will also match that sequence. For
-example, Unicode says that the letter C<LATIN SMALL LETTER SHARP S>
-should match the sequence C<ss> under C</i> rules. Thus,
+will also match that sequence. For example, Unicode says that the
+letter C<LATIN SMALL LETTER SHARP S> should match the sequence C<ss>
+under C</i> rules. Thus,
'ss' =~ /\A\N{LATIN SMALL LETTER SHARP S}\z/i # Matches
'ss' =~ /\A[aeioust\N{LATIN SMALL LETTER SHARP S}]\z/i # Matches
+For this to happen, the class must not be inverted (see L</Negation>)
+and the character must be explicitly specified, and not be part of a
+multi-character range (not even as one of its endpoints). (L</Character
+Ranges> will be explained shortly.) Therefore,
+
+ 'ss' =~ /\A[\0-\x{ff}]\z/ui # Doesn't match
+ 'ss' =~ /\A[\0-\N{LATIN SMALL LETTER SHARP S}]\z/ui # No match
+ 'ss' =~ /\A[\xDF-\xDF]\z/ui # Matches on ASCII platforms, since
+ # \xDF is LATIN SMALL LETTER SHARP S,
+ # and the range is just a single
+ # element
+
+Note that it isn't a good idea to specify these types of ranges anyway.
+
+=item *
+
+Some names known to C<\N{...}> refer to a sequence of multiple characters,
+instead of the usual single character. When one of these is included in
+the class, the entire sequence is matched. For example,
+
+ "\N{TAMIL LETTER KA}\N{TAMIL VOWEL SIGN AU}"
+ =~ / ^ [\N{TAMIL SYLLABLE KAU}] $ /x;
+
+matches, because C<\N{TAMIL SYLLABLE KAU}> is a named sequence
+consisting of the two characters matched against. Like the other
+instance where a bracketed class can match multiple characters, and for
+similar reasons, the class must not be inverted, and the named sequence
+may not appear in a range, even one where it is both endpoints. If
+these happen, it is a fatal error if the character class is within an
+extended L<C<(?[...])>|/Extended Bracketed Character Classes>
+class; and only the first code point is used (with
+a C<regexp>-type warning raised) otherwise.
+
+=back
+
=head3 Special Characters Inside a Bracketed Character Class
Most characters that are meta characters in regular expressions (that
and
C<\x>
are also special and have the same meanings as they do outside a
-bracketed character class. (However, inside a bracketed character
-class, if C<\N{I<NAME>}> expands to a sequence of characters, only the first
-one in the sequence is used, with a warning.)
+bracketed character class.
Also, a backslash followed by two or three octal digits is considered an octal
number.
"+" =~ /[+?*]/ # Match, "+" in a character class is not special.
"\cH" =~ /[\b]/ # Match, \b inside in a character class
# is equivalent to a backspace.
- "]" =~ /[][]/ # Match, as the character class contains.
+ "]" =~ /[][]/ # Match, as the character class contains
# both [ and ].
"[]" =~ /[[]]/ # Match, the pattern contains a character class
- # containing just ], and the character class is
+ # containing just [, and the character class is
# followed by a ].
=head3 Character Ranges
# hyphen ('-'), or the letter 'm'.
['-?] # Matches any of the characters '()*+,-./0123456789:;<=>?
# (But not on an EBCDIC platform).
-
+ [\N{APOSTROPHE}-\N{QUESTION MARK}]
+ # Matches any of the characters '()*+,-./0123456789:;<=>?
+ # even on an EBCDIC platform.
+ [\N{U+27}-\N{U+3F}] # Same. (U+27 is "'", and U+3F is "?")
+
+As the final two examples above show, you can achieve portablity to
+non-ASCII platforms by using the C<\N{...}> form for the range
+endpoints. These indicate that the specified range is to be interpreted
+using Unicode values, so C<[\N{U+27}-\N{U+3F}]> means to match
+C<\N{U+27}>, C<\N{U+28}>, C<\N{U+29}>, ..., C<\N{U+3D}>, C<\N{U+3E}>,
+and C<\N{U+3F}>, whatever the native code point versions for those are.
+These are called "Unicode" ranges. If either end is of the C<\N{...}>
+form, the range is considered Unicode. A C<regexp> warning is raised
+under C<S<"use re 'strict'">> if the other endpoint is specified
+non-portably:
+
+ [\N{U+00}-\x09] # Warning under re 'strict'; \x09 is non-portable
+ [\N{U+00}-\t] # No warning;
+
+Both of the above match the characters C<\N{U+00}> C<\N{U+01}>, ...
+C<\N{U+08}>, C<\N{U+09}>, but the C<\x09> looks like it could be a
+mistake so the warning is raised (under C<re 'strict'>) for it.
+
+Perl also guarantees that the ranges C<A-Z>, C<a-z>, C<0-9>, and any
+subranges of these match what an English-only speaker would expect them
+to match on any platform. That is, C<[A-Z]> matches the 26 ASCII
+uppercase letters;
+C<[a-z]> matches the 26 lowercase letters; and C<[0-9]> matches the 10
+digits. Subranges, like C<[h-k]>, match correspondingly, in this case
+just the four letters C<"h">, C<"i">, C<"j">, and C<"k">. This is the
+natural behavior on ASCII platforms where the code points (ordinal
+values) for C<"h"> through C<"k"> are consecutive integers (0x68 through
+0x6B). But special handling to achieve this may be needed on platforms
+with a non-ASCII native character set. For example, on EBCDIC
+platforms, the code point for C<"h"> is 0x88, C<"i"> is 0x89, C<"j"> is
+0x91, and C<"k"> is 0x92. Perl specially treats C<[h-k]> to exclude the
+seven code points in the gap: 0x8A through 0x90. This special handling is
+only invoked when the range is a subrange of one of the ASCII uppercase,
+lowercase, and digit ranges, AND each end of the range is expressed
+either as a literal, like C<"A">, or as a named character (C<\N{...}>,
+including the C<\N{U+...> form).
+
+EBCDIC Examples:
+
+ [i-j] # Matches either "i" or "j"
+ [i-\N{LATIN SMALL LETTER J}] # Same
+ [i-\N{U+6A}] # Same
+ [\N{U+69}-\N{U+6A}] # Same
+ [\x{89}-\x{91}] # Matches 0x89 ("i"), 0x8A .. 0x90, 0x91 ("j")
+ [i-\x{91}] # Same
+ [\x{89}-j] # Same
+ [i-J] # Matches, 0x89 ("i") .. 0xC1 ("J"); special
+ # handling doesn't apply because range is mixed
+ # case
=head3 Negation
else don't list it first.
In inverted bracketed character classes, Perl ignores the Unicode rules
-that normally say that certain characters should match a sequence of
-multiple characters under caseless C</i> matching. Following those
-rules could lead to highly confusing situations:
+that normally say that named sequence, and certain characters should
+match a sequence of multiple characters use under caseless C</i>
+matching. Following those rules could lead to highly confusing
+situations:
"ss" =~ /^[^\xDF]+$/ui; # Matches!
says that C<"ss"> is what C<\xDF> matches under C</i>. So which one
"wins"? Do you fail the match because the string has C<ss> or accept it
because it has an C<s> followed by another C<s>? Perl has chosen the
-latter.
+latter. (See note in L</Bracketed Character Classes> above.)
Examples:
X<alpha> X<alnum> X<ascii> X<blank> X<cntrl> X<digit> X<graph>
X<lower> X<print> X<punct> X<space> X<upper> X<word> X<xdigit>
-POSIX character classes have the form C<[:class:]>, where I<class> is
+POSIX character classes have the form C<[:class:]>, where I<class> is the
name, and the C<[:> and C<:]> delimiters. POSIX character classes only appear
I<inside> bracketed character classes, and are a convenient and descriptive
way of listing a group of characters.
The latter pattern would be a character class consisting of a colon,
and the letters C<a>, C<l>, C<p> and C<h>.
+
POSIX character classes can be part of a larger bracketed character class.
For example,
Perl recognizes the following POSIX character classes:
alpha Any alphabetical character ("[A-Za-z]").
- alnum Any alphanumeric character. ("[A-Za-z0-9]")
+ alnum Any alphanumeric character ("[A-Za-z0-9]").
ascii Any character in the ASCII character set.
blank A GNU extension, equal to a space or a horizontal tab ("\t").
cntrl Any control character. See Note [2] below.
lower Any lowercase character ("[a-z]").
print Any printable character, including a space. See Note [4] below.
punct Any graphical character excluding "word" characters. Note [5].
- space Any whitespace character. "\s" plus the vertical tab ("\cK").
+ space Any whitespace character. "\s" including the vertical tab
+ ("\cK").
upper Any uppercase character ("[A-Z]").
word A Perl extension ("[A-Za-z0-9_]"), equivalent to "\w".
xdigit Any hexadecimal digit ("[0-9a-fA-F]").
+Like the L<Unicode properties|/Unicode Properties>, most of the POSIX
+properties match the same regardless of whether case-insensitive (C</i>)
+matching is in effect or not. The two exceptions are C<[:upper:]> and
+C<[:lower:]>. Under C</i>, they each match the union of C<[:upper:]> and
+C<[:lower:]>.
+
Most POSIX character classes have two Unicode-style C<\p> property
counterparts. (They are not official Unicode properties, but Perl extensions
derived from official Unicode properties.) The table below shows the relation
Control characters don't produce output as such, but instead usually control
the terminal somehow: for example, newline and backspace are control characters.
-In the ASCII range, characters whose code points are between 0 and 31 inclusive,
-plus 127 (C<DEL>) are control characters.
-
-On EBCDIC platforms, it is likely that the code page will define C<[[:cntrl:]]>
-to be the EBCDIC equivalents of the ASCII controls, plus the controls
-that in Unicode have code points from 128 through 159.
+On ASCII platforms, in the ASCII range, characters whose code points are
+between 0 and 31 inclusive, plus 127 (C<DEL>) are control characters; on
+EBCDIC platforms, their counterparts are control characters.
=item [3]
=item [6]
-C<\p{SpacePerl}> and C<\p{Space}> differ only in that in non-locale
-matching, C<\p{Space}> additionally
-matches the vertical tab, C<\cK>. Same for the two ASCII-only range forms.
+C<\p{XPerlSpace}> and C<\p{Space}> match identically starting with Perl
+v5.18. In earlier versions, these differ only in that in non-locale
+matching, C<\p{XPerlSpace}> did not match the vertical tab, C<\cK>.
+Same for the two ASCII-only range forms.
=back
There are various other synonyms that can be used besides the names
-listed in the table. For example, C<\p{PosixAlpha}> can be written as
+listed in the table. For example, C<\p{XPosixAlpha}> can be written as
C<\p{Alpha}>. All are listed in
-L<perluniprops/Properties accessible through \p{} and \P{}>,
-plus all characters matched by each ASCII-range property.
+L<perluniprops/Properties accessible through \p{} and \P{}>.
Both the C<\p> counterparts always assume Unicode rules are in effect.
On ASCII platforms, this means they assume that the code points from 128
=item if locale rules are in effect ...
-The POSIX class matches according to the locale, except that
-C<word> uses the platform's native underscore character, no matter what
+The POSIX class matches according to the locale, except:
+
+=over
+
+=item C<word>
+
+also includes the platform's native underscore character, no matter what
the locale is.
-=item if Unicode rules are in effect or if on an EBCDIC platform ...
+=item C<ascii>
+
+on platforms that don't have the POSIX C<ascii> extension, this matches
+just the platform's native ASCII-range characters.
+
+=item C<blank>
+
+on platforms that don't have the POSIX C<blank> extension, this matches
+just the platform's native tab and space characters.
+
+=back
+
+=item if Unicode rules are in effect ...
The POSIX class matches the same as the Full-range counterpart.
It is proposed to change this behavior in a future release of Perl so that
whether or not Unicode rules are in effect would not change the
-behavior: Outside of locale or an EBCDIC code page, the POSIX classes
+behavior: Outside of locale, the POSIX classes
would behave like their ASCII-range counterparts. If you wish to
comment on this proposal, send email to C<perl5-porters@perl.org>.
# matches anything that isn't a hex digit.
# The OR adds the digits, leaving only the
# letters 'a' to 'f' and 'A' to 'F' excluded.
+
+=head3 Extended Bracketed Character Classes
+X<character class>
+X<set operations>
+
+This is a fancy bracketed character class that can be used for more
+readable and less error-prone classes, and to perform set operations,
+such as intersection. An example is
+
+ /(?[ \p{Thai} & \p{Digit} ])/
+
+This will match all the digit characters that are in the Thai script.
+
+This is an experimental feature available starting in 5.18, and is
+subject to change as we gain field experience with it. Any attempt to
+use it will raise a warning, unless disabled via
+
+ no warnings "experimental::regex_sets";
+
+Comments on this feature are welcome; send email to
+C<perl5-porters@perl.org>.
+
+The rules used by L<C<use re 'strict>|re/'strict' mode> apply to this
+construct.
+
+We can extend the example above:
+
+ /(?[ ( \p{Thai} + \p{Lao} ) & \p{Digit} ])/
+
+This matches digits that are in either the Thai or Laotian scripts.
+
+Notice the white space in these examples. This construct always has
+the C<E<sol>x> modifier turned on within it.
+
+The available binary operators are:
+
+ & intersection
+ + union
+ | another name for '+', hence means union
+ - subtraction (the result matches the set consisting of those
+ code points matched by the first operand, excluding any that
+ are also matched by the second operand)
+ ^ symmetric difference (the union minus the intersection). This
+ is like an exclusive or, in that the result is the set of code
+ points that are matched by either, but not both, of the
+ operands.
+
+There is one unary operator:
+
+ ! complement
+
+All the binary operators left associate; C<"&"> is higher precedence
+than the others, which all have equal precedence. The unary operator
+right associates, and has highest precedence. Thus this follows the
+normal Perl precedence rules for logical operators. Use parentheses to
+override the default precedence and associativity.
+
+The main restriction is that everything is a metacharacter. Thus,
+you cannot refer to single characters by doing something like this:
+
+ /(?[ a + b ])/ # Syntax error!
+
+The easiest way to specify an individual typable character is to enclose
+it in brackets:
+
+ /(?[ [a] + [b] ])/
+
+(This is the same thing as C<[ab]>.) You could also have said the
+equivalent:
+
+ /(?[[ a b ]])/
+
+(You can, of course, specify single characters by using, C<\x{...}>,
+C<\N{...}>, etc.)
+
+This last example shows the use of this construct to specify an ordinary
+bracketed character class without additional set operations. Note the
+white space within it; a limited version of C<E<sol>x> is turned on even
+within bracketed character classes, with only the SPACE and TAB (C<\t>)
+characters allowed, and no comments. Hence,
+
+ (?[ [#] ])
+
+matches the literal character "#". To specify a literal white space character,
+you can escape it with a backslash, like:
+
+ /(?[ [ a e i o u \ ] ])/
+
+This matches the English vowels plus the SPACE character.
+All the other escapes accepted by normal bracketed character classes are
+accepted here as well; but unrecognized escapes that generate warnings
+in normal classes are fatal errors here.
+
+All warnings from these class elements are fatal, as well as some
+practices that don't currently warn. For example you cannot say
+
+ /(?[ [ \xF ] ])/ # Syntax error!
+
+You have to have two hex digits after a braceless C<\x> (use a leading
+zero to make two). These restrictions are to lower the incidence of
+typos causing the class to not match what you thought it would.
+
+If a regular bracketed character class contains a C<\p{}> or C<\P{}> and
+is matched against a non-Unicode code point, a warning may be
+raised, as the result is not Unicode-defined. No such warning will come
+when using this extended form.
+
+The final difference between regular bracketed character classes and
+these, is that it is not possible to get these to match a
+multi-character fold. Thus,
+
+ /(?[ [\xDF] ])/iu
+
+does not match the string C<ss>.
+
+You don't have to enclose POSIX class names inside double brackets,
+hence both of the following work:
+
+ /(?[ [:word:] - [:lower:] ])/
+ /(?[ [[:word:]] - [[:lower:]] ])/
+
+Any contained POSIX character classes, including things like C<\w> and C<\D>
+respect the C<E<sol>a> (and C<E<sol>aa>) modifiers.
+
+C<< (?[ ]) >> is a regex-compile-time construct. Any attempt to use
+something which isn't knowable at the time the containing regular
+expression is compiled is a fatal error. In practice, this means
+just three limitations:
+
+=over 4
+
+=item 1
+
+When compiled within the scope of C<use locale> (or the C<E<sol>l> regex
+modifier), this construct assumes that the execution-time locale will be
+a UTF-8 one, and the generated pattern always uses Unicode rules. What
+gets matched or not thus isn't dependent on the actual runtime locale, so
+tainting is not enabled. But a C<locale> category warning is raised
+if the runtime locale turns out to not be UTF-8.
+
+=item 2
+
+Any
+L<user-defined property|perlunicode/"User-Defined Character Properties">
+used must be already defined by the time the regular expression is
+compiled (but note that this construct can be used instead of such
+properties).
+
+=item 3
+
+A regular expression that otherwise would compile
+using C<E<sol>d> rules, and which uses this construct will instead
+use C<E<sol>u>. Thus this construct tells Perl that you don't want
+C<E<sol>d> rules for the entire regular expression containing it.
+
+=back
+
+Note that skipping white space applies only to the interior of this
+construct. There must not be any space between any of the characters
+that form the initial C<(?[>. Nor may there be space between the
+closing C<])> characters.
+
+Just as in all regular expressions, the pattern can be built up by
+including variables that are interpolated at regex compilation time.
+Care must be taken to ensure that you are getting what you expect. For
+example:
+
+ my $thai_or_lao = '\p{Thai} + \p{Lao}';
+ ...
+ qr/(?[ \p{Digit} & $thai_or_lao ])/;
+
+compiles to
+
+ qr/(?[ \p{Digit} & \p{Thai} + \p{Lao} ])/;
+
+But this does not have the effect that someone reading the code would
+likely expect, as the intersection applies just to C<\p{Thai}>,
+excluding the Laotian. Pitfalls like this can be avoided by
+parenthesizing the component pieces:
+
+ my $thai_or_lao = '( \p{Thai} + \p{Lao} )';
+
+But any modifiers will still apply to all the components:
+
+ my $lower = '\p{Lower} + \p{Digit}';
+ qr/(?[ \p{Greek} & $lower ])/i;
+
+matches upper case things. You can avoid surprises by making the
+components into instances of this construct by compiling them:
+
+ my $thai_or_lao = qr/(?[ \p{Thai} + \p{Lao} ])/;
+ my $lower = qr/(?[ \p{Lower} + \p{Digit} ])/;
+
+When these are embedded in another pattern, what they match does not
+change, regardless of parenthesization or what modifiers are in effect
+in that outer pattern.
+
+Due to the way that Perl parses things, your parentheses and brackets
+may need to be balanced, even including comments. If you run into any
+examples, please send them to C<perlbug@perl.org>, so that we can have a
+concrete example for this man page.
+
+We may change it so that things that remain legal uses in normal bracketed
+character classes might become illegal within this experimental
+construct. One proposal, for example, is to forbid adjacent uses of the
+same character, as in C<(?[ [aa] ])>. The motivation for such a change
+is that this usage is likely a typo, as the second "a" adds nothing.