3 perlunicode - Unicode support in Perl
7 =head2 Important Caveats
9 Unicode support is an extensive requirement. While Perl does not
10 implement the Unicode standard or the accompanying technical reports
11 from cover to cover, Perl does support many Unicode features.
13 People who want to learn to use Unicode in Perl, should probably read
14 L<the Perl Unicode tutorial|perlunitut> before reading this reference
19 =item Input and Output Layers
21 Perl knows when a filehandle uses Perl's internal Unicode encodings
22 (UTF-8, or UTF-EBCDIC if in EBCDIC) if the filehandle is opened with
23 the ":utf8" layer. Other encodings can be converted to Perl's
24 encoding on input or from Perl's encoding on output by use of the
25 ":encoding(...)" layer. See L<open>.
27 To indicate that Perl source itself is in UTF-8, use C<use utf8;>.
29 =item Regular Expressions
31 The regular expression compiler produces polymorphic opcodes. That is,
32 the pattern adapts to the data and automatically switches to the Unicode
33 character scheme when presented with data that is internally encoded in
34 UTF-8 -- or instead uses a traditional byte scheme when presented with
37 =item C<use utf8> still needed to enable UTF-8/UTF-EBCDIC in scripts
39 As a compatibility measure, the C<use utf8> pragma must be explicitly
40 included to enable recognition of UTF-8 in the Perl scripts themselves
41 (in string or regular expression literals, or in identifier names) on
42 ASCII-based machines or to recognize UTF-EBCDIC on EBCDIC-based
43 machines. B<These are the only times when an explicit C<use utf8>
44 is needed.> See L<utf8>.
46 =item BOM-marked scripts and UTF-16 scripts autodetected
48 If a Perl script begins marked with the Unicode BOM (UTF-16LE, UTF16-BE,
49 or UTF-8), or if the script looks like non-BOM-marked UTF-16 of either
50 endianness, Perl will correctly read in the script as Unicode.
51 (BOMless UTF-8 cannot be effectively recognized or differentiated from
52 ISO 8859-1 or other eight-bit encodings.)
54 =item C<use encoding> needed to upgrade non-Latin-1 byte strings
56 By default, there is a fundamental asymmetry in Perl's Unicode model:
57 implicit upgrading from byte strings to Unicode strings assumes that
58 they were encoded in I<ISO 8859-1 (Latin-1)>, but Unicode strings are
59 downgraded with UTF-8 encoding. This happens because the first 256
60 codepoints in Unicode happens to agree with Latin-1.
62 See L</"Byte and Character Semantics"> for more details.
66 =head2 Byte and Character Semantics
68 Beginning with version 5.6, Perl uses logically-wide characters to
69 represent strings internally.
71 In future, Perl-level operations will be expected to work with
72 characters rather than bytes.
74 However, as an interim compatibility measure, Perl aims to
75 provide a safe migration path from byte semantics to character
76 semantics for programs. For operations where Perl can unambiguously
77 decide that the input data are characters, Perl switches to
78 character semantics. For operations where this determination cannot
79 be made without additional information from the user, Perl decides in
80 favor of compatibility and chooses to use byte semantics.
82 This behavior preserves compatibility with earlier versions of Perl,
83 which allowed byte semantics in Perl operations only if
84 none of the program's inputs were marked as being as source of Unicode
85 character data. Such data may come from filehandles, from calls to
86 external programs, from information provided by the system (such as %ENV),
87 or from literals and constants in the source text.
89 The C<bytes> pragma will always, regardless of platform, force byte
90 semantics in a particular lexical scope. See L<bytes>.
92 The C<utf8> pragma is primarily a compatibility device that enables
93 recognition of UTF-(8|EBCDIC) in literals encountered by the parser.
94 Note that this pragma is only required while Perl defaults to byte
95 semantics; when character semantics become the default, this pragma
96 may become a no-op. See L<utf8>.
98 Unless explicitly stated, Perl operators use character semantics
99 for Unicode data and byte semantics for non-Unicode data.
100 The decision to use character semantics is made transparently. If
101 input data comes from a Unicode source--for example, if a character
102 encoding layer is added to a filehandle or a literal Unicode
103 string constant appears in a program--character semantics apply.
104 Otherwise, byte semantics are in effect. The C<bytes> pragma should
105 be used to force byte semantics on Unicode data.
107 If strings operating under byte semantics and strings with Unicode
108 character data are concatenated, the new string will be created by
109 decoding the byte strings as I<ISO 8859-1 (Latin-1)>, even if the
110 old Unicode string used EBCDIC. This translation is done without
111 regard to the system's native 8-bit encoding.
113 Under character semantics, many operations that formerly operated on
114 bytes now operate on characters. A character in Perl is
115 logically just a number ranging from 0 to 2**31 or so. Larger
116 characters may encode into longer sequences of bytes internally, but
117 this internal detail is mostly hidden for Perl code.
118 See L<perluniintro> for more.
120 =head2 Effects of Character Semantics
122 Character semantics have the following effects:
128 Strings--including hash keys--and regular expression patterns may
129 contain characters that have an ordinal value larger than 255.
131 If you use a Unicode editor to edit your program, Unicode characters may
132 occur directly within the literal strings in UTF-8 encoding, or UTF-16.
133 (The former requires a BOM or C<use utf8>, the latter requires a BOM.)
135 Unicode characters can also be added to a string by using the C<\x{...}>
136 notation. The Unicode code for the desired character, in hexadecimal,
137 should be placed in the braces. For instance, a smiley face is
138 C<\x{263A}>. This encoding scheme only works for all characters, but
139 for characters under 0x100, note that Perl may use an 8 bit encoding
140 internally, for optimization and/or backward compatibility.
144 use charnames ':full';
146 you can use the C<\N{...}> notation and put the official Unicode
147 character name within the braces, such as C<\N{WHITE SMILING FACE}>.
151 If an appropriate L<encoding> is specified, identifiers within the
152 Perl script may contain Unicode alphanumeric characters, including
153 ideographs. Perl does not currently attempt to canonicalize variable
158 Regular expressions match characters instead of bytes. "." matches
159 a character instead of a byte.
163 Character classes in regular expressions match characters instead of
164 bytes and match against the character properties specified in the
165 Unicode properties database. C<\w> can be used to match a Japanese
166 ideograph, for instance.
170 Named Unicode properties, scripts, and block ranges may be used like
171 character classes via the C<\p{}> "matches property" construct and
172 the C<\P{}> negation, "doesn't match property".
174 See L</"Unicode Character Properties"> for more details.
176 You can define your own character properties and use them
177 in the regular expression with the C<\p{}> or C<\P{}> construct.
179 See L</"User-Defined Character Properties"> for more details.
183 The special pattern C<\X> matches any extended Unicode
184 sequence--"a combining character sequence" in Standardese--where the
185 first character is a base character and subsequent characters are mark
186 characters that apply to the base character. C<\X> is equivalent to
191 The C<tr///> operator translates characters instead of bytes. Note
192 that the C<tr///CU> functionality has been removed. For similar
193 functionality see pack('U0', ...) and pack('C0', ...).
197 Case translation operators use the Unicode case translation tables
198 when character input is provided. Note that C<uc()>, or C<\U> in
199 interpolated strings, translates to uppercase, while C<ucfirst>,
200 or C<\u> in interpolated strings, translates to titlecase in languages
201 that make the distinction.
205 Most operators that deal with positions or lengths in a string will
206 automatically switch to using character positions, including
207 C<chop()>, C<chomp()>, C<substr()>, C<pos()>, C<index()>, C<rindex()>,
208 C<sprintf()>, C<write()>, and C<length()>. An operator that
209 specifically does not switch is C<vec()>. Operators that really don't
210 care include operators that treat strings as a bucket of bits such as
211 C<sort()>, and operators dealing with filenames.
215 The C<pack()>/C<unpack()> letter C<C> does I<not> change, since it is often
216 used for byte-oriented formats. Again, think C<char> in the C language.
218 There is a new C<U> specifier that converts between Unicode characters
219 and code points. There is also a C<W> specifier that is the equivalent of
220 C<chr>/C<ord> and properly handles character values even if they are above 255.
224 The C<chr()> and C<ord()> functions work on characters, similar to
225 C<pack("W")> and C<unpack("W")>, I<not> C<pack("C")> and
226 C<unpack("C")>. C<pack("C")> and C<unpack("C")> are methods for
227 emulating byte-oriented C<chr()> and C<ord()> on Unicode strings.
228 While these methods reveal the internal encoding of Unicode strings,
229 that is not something one normally needs to care about at all.
233 The bit string operators, C<& | ^ ~>, can operate on character data.
234 However, for backward compatibility, such as when using bit string
235 operations when characters are all less than 256 in ordinal value, one
236 should not use C<~> (the bit complement) with characters of both
237 values less than 256 and values greater than 256. Most importantly,
238 DeMorgan's laws (C<~($x|$y) eq ~$x&~$y> and C<~($x&$y) eq ~$x|~$y>)
239 will not hold. The reason for this mathematical I<faux pas> is that
240 the complement cannot return B<both> the 8-bit (byte-wide) bit
241 complement B<and> the full character-wide bit complement.
245 lc(), uc(), lcfirst(), and ucfirst() work for the following cases:
251 the case mapping is from a single Unicode character to another
252 single Unicode character, or
256 the case mapping is from a single Unicode character to more
257 than one Unicode character.
261 Things to do with locales (Lithuanian, Turkish, Azeri) do B<not> work
262 since Perl does not understand the concept of Unicode locales.
264 See the Unicode Technical Report #21, Case Mappings, for more details.
266 But you can also define your own mappings to be used in the lc(),
267 lcfirst(), uc(), and ucfirst() (or their string-inlined versions).
269 See L</"User-Defined Case Mappings"> for more details.
277 And finally, C<scalar reverse()> reverses by character rather than by byte.
281 =head2 Unicode Character Properties
283 Named Unicode properties, scripts, and block ranges may be used like
284 character classes via the C<\p{}> "matches property" construct and
285 the C<\P{}> negation, "doesn't match property".
287 For instance, C<\p{Lu}> matches any character with the Unicode "Lu"
288 (Letter, uppercase) property, while C<\p{M}> matches any character
289 with an "M" (mark--accents and such) property. Brackets are not
290 required for single letter properties, so C<\p{M}> is equivalent to
291 C<\pM>. Many predefined properties are available, such as
292 C<\p{Mirrored}> and C<\p{Tibetan}>.
294 The official Unicode script and block names have spaces and dashes as
295 separators, but for convenience you can use dashes, spaces, or
296 underbars, and case is unimportant. It is recommended, however, that
297 for consistency you use the following naming: the official Unicode
298 script, property, or block name (see below for the additional rules
299 that apply to block names) with whitespace and dashes removed, and the
300 words "uppercase-first-lowercase-rest". C<Latin-1 Supplement> thus
301 becomes C<Latin1Supplement>.
303 You can also use negation in both C<\p{}> and C<\P{}> by introducing a caret
304 (^) between the first brace and the property name: C<\p{^Tamil}> is
305 equal to C<\P{Tamil}>.
307 B<NOTE: the properties, scripts, and blocks listed here are as of
308 Unicode 5.0.0 in July 2006.>
312 =item General Category
314 Here are the basic Unicode General Category properties, followed by their
315 long form. You can use either; C<\p{Lu}> and C<\p{UppercaseLetter}>,
316 for instance, are identical.
339 Pc ConnectorPunctuation
343 Pi InitialPunctuation
344 (may behave like Ps or Pe depending on usage)
346 (may behave like Ps or Pe depending on usage)
358 Zp ParagraphSeparator
363 Cs Surrogate (not usable)
367 Single-letter properties match all characters in any of the
368 two-letter sub-properties starting with the same letter.
369 C<LC> and C<L&> are special cases, which are aliases for the set of
370 C<Ll>, C<Lu>, and C<Lt>.
372 Because Perl hides the need for the user to understand the internal
373 representation of Unicode characters, there is no need to implement
374 the somewhat messy concept of surrogates. C<Cs> is therefore not
377 =item Bidirectional Character Types
379 Because scripts differ in their directionality--Hebrew is
380 written right to left, for example--Unicode supplies these properties in
386 LRE Left-to-Right Embedding
387 LRO Left-to-Right Override
389 AL Right-to-Left Arabic
390 RLE Right-to-Left Embedding
391 RLO Right-to-Left Override
392 PDF Pop Directional Format
394 ES European Number Separator
395 ET European Number Terminator
397 CS Common Number Separator
400 B Paragraph Separator
405 For example, C<\p{BidiClass:R}> matches characters that are normally
406 written right to left.
410 The script names which can be used by C<\p{...}> and C<\P{...}>,
411 such as in C<\p{Latin}> or C<\p{Cyrillic}>, are as follows:
479 =item Extended property classes
481 Extended property classes can supplement the basic
482 properties, defined by the F<PropList> Unicode database:
496 LogicalOrderException
497 NoncharacterCodePoint
499 OtherDefaultIgnorableCodePoint
517 and there are further derived properties:
519 Alphabetic = Lu + Ll + Lt + Lm + Lo + Nl + OtherAlphabetic
520 Lowercase = Ll + OtherLowercase
521 Uppercase = Lu + OtherUppercase
522 Math = Sm + OtherMath
524 IDStart = Lu + Ll + Lt + Lm + Lo + Nl + OtherIDStart
525 IDContinue = IDStart + Mn + Mc + Nd + Pc + OtherIDContinue
527 DefaultIgnorableCodePoint
528 = OtherDefaultIgnorableCodePoint
529 + Cf + Cc + Cs + Noncharacters + VariationSelector
530 - WhiteSpace - FFF9..FFFB (Annotation Characters)
532 Any = Any code points (i.e. U+0000 to U+10FFFF)
533 Assigned = Any non-Cn code points (i.e. synonym for \P{Cn})
534 Unassigned = Synonym for \p{Cn}
535 ASCII = ASCII (i.e. U+0000 to U+007F)
537 Common = Any character (or unassigned code point)
538 not explicitly assigned to a script
540 =item Use of "Is" Prefix
542 For backward compatibility (with Perl 5.6), all properties mentioned
543 so far may have C<Is> prepended to their name, so C<\P{IsLu}>, for
544 example, is equal to C<\P{Lu}>.
548 In addition to B<scripts>, Unicode also defines B<blocks> of
549 characters. The difference between scripts and blocks is that the
550 concept of scripts is closer to natural languages, while the concept
551 of blocks is more of an artificial grouping based on groups of 256
552 Unicode characters. For example, the C<Latin> script contains letters
553 from many blocks but does not contain all the characters from those
554 blocks. It does not, for example, contain digits, because digits are
555 shared across many scripts. Digits and similar groups, like
556 punctuation, are in a category called C<Common>.
558 For more about scripts, see the UAX#24 "Script Names":
560 http://www.unicode.org/reports/tr24/
562 For more about blocks, see:
564 http://www.unicode.org/Public/UNIDATA/Blocks.txt
566 Block names are given with the C<In> prefix. For example, the
567 Katakana block is referenced via C<\p{InKatakana}>. The C<In>
568 prefix may be omitted if there is no naming conflict with a script
569 or any other property, but it is recommended that C<In> always be used
570 for block tests to avoid confusion.
572 These block names are supported:
575 InAlphabeticPresentationForms
576 InAncientGreekMusicalNotation
577 InAncientGreekNumbers
579 InArabicPresentationFormsA
580 InArabicPresentationFormsB
594 InByzantineMusicalSymbols
596 InCJKCompatibilityForms
597 InCJKCompatibilityIdeographs
598 InCJKCompatibilityIdeographsSupplement
599 InCJKRadicalsSupplement
601 InCJKSymbolsAndPunctuation
602 InCJKUnifiedIdeographs
603 InCJKUnifiedIdeographsExtensionA
604 InCJKUnifiedIdeographsExtensionB
606 InCombiningDiacriticalMarks
607 InCombiningDiacriticalMarksSupplement
608 InCombiningDiacriticalMarksforSymbols
612 InCountingRodNumerals
614 InCuneiformNumbersAndPunctuation
622 InEnclosedAlphanumerics
623 InEnclosedCJKLettersAndMonths
637 InHalfwidthAndFullwidthForms
638 InHangulCompatibilityJamo
643 InHighPrivateUseSurrogates
647 InIdeographicDescriptionCharacters
652 InKatakanaPhoneticExtensions
659 InLatinExtendedAdditional
669 InMathematicalAlphanumericSymbols
670 InMathematicalOperators
671 InMiscellaneousMathematicalSymbolsA
672 InMiscellaneousMathematicalSymbolsB
673 InMiscellaneousSymbols
674 InMiscellaneousSymbolsAndArrows
675 InMiscellaneousTechnical
676 InModifierToneLetters
686 InOpticalCharacterRecognition
692 InPhoneticExtensionsSupplement
698 InSpacingModifierLetters
700 InSuperscriptsAndSubscripts
701 InSupplementalArrowsA
702 InSupplementalArrowsB
703 InSupplementalMathematicalOperators
704 InSupplementalPunctuation
705 InSupplementaryPrivateUseAreaA
706 InSupplementaryPrivateUseAreaB
721 InUnifiedCanadianAboriginalSyllabics
723 InVariationSelectorsSupplement
727 InYijingHexagramSymbols
731 =head2 User-Defined Character Properties
733 You can define your own character properties by defining subroutines
734 whose names begin with "In" or "Is". The subroutines can be defined in
735 any package. The user-defined properties can be used in the regular
736 expression C<\p> and C<\P> constructs; if you are using a user-defined
737 property from a package other than the one you are in, you must specify
738 its package in the C<\p> or C<\P> construct.
740 # assuming property IsForeign defined in Lang::
741 package main; # property package name required
742 if ($txt =~ /\p{Lang::IsForeign}+/) { ... }
744 package Lang; # property package name not required
745 if ($txt =~ /\p{IsForeign}+/) { ... }
748 Note that the effect is compile-time and immutable once defined.
750 The subroutines must return a specially-formatted string, with one
751 or more newline-separated lines. Each line must be one of the following:
757 A single hexadecimal number denoting a Unicode code point to include.
761 Two hexadecimal numbers separated by horizontal whitespace (space or
762 tabular characters) denoting a range of Unicode code points to include.
766 Something to include, prefixed by "+": a built-in character
767 property (prefixed by "utf8::") or a user-defined character property,
768 to represent all the characters in that property; two hexadecimal code
769 points for a range; or a single hexadecimal code point.
773 Something to exclude, prefixed by "-": an existing character
774 property (prefixed by "utf8::") or a user-defined character property,
775 to represent all the characters in that property; two hexadecimal code
776 points for a range; or a single hexadecimal code point.
780 Something to negate, prefixed "!": an existing character
781 property (prefixed by "utf8::") or a user-defined character property,
782 to represent all the characters in that property; two hexadecimal code
783 points for a range; or a single hexadecimal code point.
787 Something to intersect with, prefixed by "&": an existing character
788 property (prefixed by "utf8::") or a user-defined character property,
789 for all the characters except the characters in the property; two
790 hexadecimal code points for a range; or a single hexadecimal code point.
794 For example, to define a property that covers both the Japanese
795 syllabaries (hiragana and katakana), you can define
804 Imagine that the here-doc end marker is at the beginning of the line.
805 Now you can use C<\p{InKana}> and C<\P{InKana}>.
807 You could also have used the existing block property names:
816 Suppose you wanted to match only the allocated characters,
817 not the raw block ranges: in other words, you want to remove
828 The negation is useful for defining (surprise!) negated classes.
838 Intersection is useful for getting the common characters matched by
839 two (or more) classes.
848 It's important to remember not to use "&" for the first set -- that
849 would be intersecting with nothing (resulting in an empty set).
851 =head2 User-Defined Case Mappings
853 You can also define your own mappings to be used in the lc(),
854 lcfirst(), uc(), and ucfirst() (or their string-inlined versions).
855 The principle is similar to that of user-defined character
856 properties: to define subroutines in the C<main> package
857 with names like C<ToLower> (for lc() and lcfirst()), C<ToTitle> (for
858 the first character in ucfirst()), and C<ToUpper> (for uc(), and the
859 rest of the characters in ucfirst()).
861 The string returned by the subroutines needs now to be three
862 hexadecimal numbers separated by tabulators: start of the source
863 range, end of the source range, and start of the destination range.
872 defines an uc() mapping that causes only the characters "a", "b", and
873 "c" to be mapped to "A", "B", "C", all other characters will remain
876 If there is no source range to speak of, that is, the mapping is from
877 a single character to another single character, leave the end of the
878 source range empty, but the two tabulator characters are still needed.
887 defines a lc() mapping that causes only "A" to be mapped to "a", all
888 other characters will remain unchanged.
890 (For serious hackers only) If you want to introspect the default
891 mappings, you can find the data in the directory
892 C<$Config{privlib}>/F<unicore/To/>. The mapping data is returned as
893 the here-document, and the C<utf8::ToSpecFoo> are special exception
894 mappings derived from <$Config{privlib}>/F<unicore/SpecialCasing.txt>.
895 The C<Digit> and C<Fold> mappings that one can see in the directory
896 are not directly user-accessible, one can use either the
897 C<Unicode::UCD> module, or just match case-insensitively (that's when
898 the C<Fold> mapping is used).
900 A final note on the user-defined case mappings: they will be used
901 only if the scalar has been marked as having Unicode characters.
902 Old byte-style strings will not be affected.
904 =head2 Character Encodings for Input and Output
908 =head2 Unicode Regular Expression Support Level
910 The following list of Unicode support for regular expressions describes
911 all the features currently supported. The references to "Level N"
912 and the section numbers refer to the Unicode Technical Standard #18,
913 "Unicode Regular Expressions", version 11, in May 2005.
919 Level 1 - Basic Unicode Support
921 RL1.1 Hex Notation - done [1]
922 RL1.2 Properties - done [2][3]
923 RL1.2a Compatibility Properties - done [4]
924 RL1.3 Subtraction and Intersection - MISSING [5]
925 RL1.4 Simple Word Boundaries - done [6]
926 RL1.5 Simple Loose Matches - done [7]
927 RL1.6 Line Boundaries - MISSING [8]
928 RL1.7 Supplementary Code Points - done [9]
932 [3] supports not only minimal list (general category, scripts,
933 Alphabetic, Lowercase, Uppercase, WhiteSpace,
934 NoncharacterCodePoint, DefaultIgnorableCodePoint, Any,
935 ASCII, Assigned), but also bidirectional types, blocks, etc.
936 (see L</"Unicode Character Properties">)
937 [4] \d \D \s \S \w \W \X [:prop:] [:^prop:]
938 [5] can use regular expression look-ahead [a] or
939 user-defined character properties [b] to emulate set operations
941 [7] note that Perl does Full case-folding in matching, not Simple:
942 for example U+1F88 is equivalent with U+1F00 U+03B9,
943 not with 1F80. This difference matters for certain Greek
944 capital letters with certain modifiers: the Full case-folding
945 decomposes the letter, while the Simple case-folding would map
946 it to a single character.
947 [8] should do ^ and $ also on U+000B (\v in C), FF (\f), CR (\r),
948 CRLF (\r\n), NEL (U+0085), LS (U+2028), and PS (U+2029);
949 should also affect <>, $., and script line numbers;
950 should not split lines within CRLF [c] (i.e. there is no empty
951 line between \r and \n)
952 [9] UTF-8/UTF-EBDDIC used in perl allows not only U+10000 to U+10FFFF
953 but also beyond U+10FFFF [d]
955 [a] You can mimic class subtraction using lookahead.
956 For example, what UTS#18 might write as
958 [{Greek}-[{UNASSIGNED}]]
960 in Perl can be written as:
962 (?!\p{Unassigned})\p{InGreekAndCoptic}
963 (?=\p{Assigned})\p{InGreekAndCoptic}
965 But in this particular example, you probably really want
969 which will match assigned characters known to be part of the Greek script.
971 Also see the Unicode::Regex::Set module, it does implement the full
972 UTS#18 grouping, intersection, union, and removal (subtraction) syntax.
974 [b] '+' for union, '-' for removal (set-difference), '&' for intersection
975 (see L</"User-Defined Character Properties">)
977 [c] Try the C<:crlf> layer (see L<PerlIO>).
979 [d] Avoid C<use warning 'utf8';> (or say C<no warning 'utf8';>) to allow
980 U+FFFF (C<\x{FFFF}>).
984 Level 2 - Extended Unicode Support
986 RL2.1 Canonical Equivalents - MISSING [10][11]
987 RL2.2 Default Grapheme Clusters - MISSING [12][13]
988 RL2.3 Default Word Boundaries - MISSING [14]
989 RL2.4 Default Loose Matches - MISSING [15]
990 RL2.5 Name Properties - MISSING [16]
991 RL2.6 Wildcard Properties - MISSING
993 [10] see UAX#15 "Unicode Normalization Forms"
994 [11] have Unicode::Normalize but not integrated to regexes
995 [12] have \X but at this level . should equal that
996 [13] UAX#29 "Text Boundaries" considers CRLF and Hangul syllable
997 clusters as a single grapheme cluster.
998 [14] see UAX#29, Word Boundaries
999 [15] see UAX#21 "Case Mappings"
1000 [16] have \N{...} but neither compute names of CJK Ideographs
1001 and Hangul Syllables nor use a loose match [e]
1003 [e] C<\N{...}> allows namespaces (see L<charnames>).
1007 Level 3 - Tailored Support
1009 RL3.1 Tailored Punctuation - MISSING
1010 RL3.2 Tailored Grapheme Clusters - MISSING [17][18]
1011 RL3.3 Tailored Word Boundaries - MISSING
1012 RL3.4 Tailored Loose Matches - MISSING
1013 RL3.5 Tailored Ranges - MISSING
1014 RL3.6 Context Matching - MISSING [19]
1015 RL3.7 Incremental Matches - MISSING
1016 ( RL3.8 Unicode Set Sharing )
1017 RL3.9 Possible Match Sets - MISSING
1018 RL3.10 Folded Matching - MISSING [20]
1019 RL3.11 Submatchers - MISSING
1021 [17] see UAX#10 "Unicode Collation Algorithms"
1022 [18] have Unicode::Collate but not integrated to regexes
1023 [19] have (?<=x) and (?=x), but look-aheads or look-behinds should see
1024 outside of the target substring
1025 [20] need insensitive matching for linguistic features other than case;
1026 for example, hiragana to katakana, wide and narrow, simplified Han
1027 to traditional Han (see UTR#30 "Character Foldings")
1031 =head2 Unicode Encodings
1033 Unicode characters are assigned to I<code points>, which are abstract
1034 numbers. To use these numbers, various encodings are needed.
1042 UTF-8 is a variable-length (1 to 6 bytes, current character allocations
1043 require 4 bytes), byte-order independent encoding. For ASCII (and we
1044 really do mean 7-bit ASCII, not another 8-bit encoding), UTF-8 is
1047 The following table is from Unicode 3.2.
1049 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
1051 U+0000..U+007F 00..7F
1052 U+0080..U+07FF C2..DF 80..BF
1053 U+0800..U+0FFF E0 A0..BF 80..BF
1054 U+1000..U+CFFF E1..EC 80..BF 80..BF
1055 U+D000..U+D7FF ED 80..9F 80..BF
1056 U+D800..U+DFFF ******* ill-formed *******
1057 U+E000..U+FFFF EE..EF 80..BF 80..BF
1058 U+10000..U+3FFFF F0 90..BF 80..BF 80..BF
1059 U+40000..U+FFFFF F1..F3 80..BF 80..BF 80..BF
1060 U+100000..U+10FFFF F4 80..8F 80..BF 80..BF
1062 Note the C<A0..BF> in C<U+0800..U+0FFF>, the C<80..9F> in
1063 C<U+D000...U+D7FF>, the C<90..B>F in C<U+10000..U+3FFFF>, and the
1064 C<80...8F> in C<U+100000..U+10FFFF>. The "gaps" are caused by legal
1065 UTF-8 avoiding non-shortest encodings: it is technically possible to
1066 UTF-8-encode a single code point in different ways, but that is
1067 explicitly forbidden, and the shortest possible encoding should always
1068 be used. So that's what Perl does.
1070 Another way to look at it is via bits:
1072 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
1075 00000bbbbbaaaaaa 110bbbbb 10aaaaaa
1076 ccccbbbbbbaaaaaa 1110cccc 10bbbbbb 10aaaaaa
1077 00000dddccccccbbbbbbaaaaaa 11110ddd 10cccccc 10bbbbbb 10aaaaaa
1079 As you can see, the continuation bytes all begin with C<10>, and the
1080 leading bits of the start byte tell how many bytes the are in the
1087 Like UTF-8 but EBCDIC-safe, in the way that UTF-8 is ASCII-safe.
1091 UTF-16, UTF-16BE, UTF-16LE, Surrogates, and BOMs (Byte Order Marks)
1093 The followings items are mostly for reference and general Unicode
1094 knowledge, Perl doesn't use these constructs internally.
1096 UTF-16 is a 2 or 4 byte encoding. The Unicode code points
1097 C<U+0000..U+FFFF> are stored in a single 16-bit unit, and the code
1098 points C<U+10000..U+10FFFF> in two 16-bit units. The latter case is
1099 using I<surrogates>, the first 16-bit unit being the I<high
1100 surrogate>, and the second being the I<low surrogate>.
1102 Surrogates are code points set aside to encode the C<U+10000..U+10FFFF>
1103 range of Unicode code points in pairs of 16-bit units. The I<high
1104 surrogates> are the range C<U+D800..U+DBFF>, and the I<low surrogates>
1105 are the range C<U+DC00..U+DFFF>. The surrogate encoding is
1107 $hi = ($uni - 0x10000) / 0x400 + 0xD800;
1108 $lo = ($uni - 0x10000) % 0x400 + 0xDC00;
1112 $uni = 0x10000 + ($hi - 0xD800) * 0x400 + ($lo - 0xDC00);
1114 If you try to generate surrogates (for example by using chr()), you
1115 will get a warning if warnings are turned on, because those code
1116 points are not valid for a Unicode character.
1118 Because of the 16-bitness, UTF-16 is byte-order dependent. UTF-16
1119 itself can be used for in-memory computations, but if storage or
1120 transfer is required either UTF-16BE (big-endian) or UTF-16LE
1121 (little-endian) encodings must be chosen.
1123 This introduces another problem: what if you just know that your data
1124 is UTF-16, but you don't know which endianness? Byte Order Marks, or
1125 BOMs, are a solution to this. A special character has been reserved
1126 in Unicode to function as a byte order marker: the character with the
1127 code point C<U+FEFF> is the BOM.
1129 The trick is that if you read a BOM, you will know the byte order,
1130 since if it was written on a big-endian platform, you will read the
1131 bytes C<0xFE 0xFF>, but if it was written on a little-endian platform,
1132 you will read the bytes C<0xFF 0xFE>. (And if the originating platform
1133 was writing in UTF-8, you will read the bytes C<0xEF 0xBB 0xBF>.)
1135 The way this trick works is that the character with the code point
1136 C<U+FFFE> is guaranteed not to be a valid Unicode character, so the
1137 sequence of bytes C<0xFF 0xFE> is unambiguously "BOM, represented in
1138 little-endian format" and cannot be C<U+FFFE>, represented in big-endian
1143 UTF-32, UTF-32BE, UTF-32LE
1145 The UTF-32 family is pretty much like the UTF-16 family, expect that
1146 the units are 32-bit, and therefore the surrogate scheme is not
1147 needed. The BOM signatures will be C<0x00 0x00 0xFE 0xFF> for BE and
1148 C<0xFF 0xFE 0x00 0x00> for LE.
1154 Encodings defined by the ISO 10646 standard. UCS-2 is a 16-bit
1155 encoding. Unlike UTF-16, UCS-2 is not extensible beyond C<U+FFFF>,
1156 because it does not use surrogates. UCS-4 is a 32-bit encoding,
1157 functionally identical to UTF-32.
1163 A seven-bit safe (non-eight-bit) encoding, which is useful if the
1164 transport or storage is not eight-bit safe. Defined by RFC 2152.
1168 =head2 Security Implications of Unicode
1176 Unfortunately, the specification of UTF-8 leaves some room for
1177 interpretation of how many bytes of encoded output one should generate
1178 from one input Unicode character. Strictly speaking, the shortest
1179 possible sequence of UTF-8 bytes should be generated,
1180 because otherwise there is potential for an input buffer overflow at
1181 the receiving end of a UTF-8 connection. Perl always generates the
1182 shortest length UTF-8, and with warnings on Perl will warn about
1183 non-shortest length UTF-8 along with other malformations, such as the
1184 surrogates, which are not real Unicode code points.
1188 Regular expressions behave slightly differently between byte data and
1189 character (Unicode) data. For example, the "word character" character
1190 class C<\w> will work differently depending on if data is eight-bit bytes
1193 In the first case, the set of C<\w> characters is either small--the
1194 default set of alphabetic characters, digits, and the "_"--or, if you
1195 are using a locale (see L<perllocale>), the C<\w> might contain a few
1196 more letters according to your language and country.
1198 In the second case, the C<\w> set of characters is much, much larger.
1199 Most importantly, even in the set of the first 256 characters, it will
1200 probably match different characters: unlike most locales, which are
1201 specific to a language and country pair, Unicode classifies all the
1202 characters that are letters I<somewhere> as C<\w>. For example, your
1203 locale might not think that LATIN SMALL LETTER ETH is a letter (unless
1204 you happen to speak Icelandic), but Unicode does.
1206 As discussed elsewhere, Perl has one foot (two hooves?) planted in
1207 each of two worlds: the old world of bytes and the new world of
1208 characters, upgrading from bytes to characters when necessary.
1209 If your legacy code does not explicitly use Unicode, no automatic
1210 switch-over to characters should happen. Characters shouldn't get
1211 downgraded to bytes, either. It is possible to accidentally mix bytes
1212 and characters, however (see L<perluniintro>), in which case C<\w> in
1213 regular expressions might start behaving differently. Review your
1214 code. Use warnings and the C<strict> pragma.
1218 =head2 Unicode in Perl on EBCDIC
1220 The way Unicode is handled on EBCDIC platforms is still
1221 experimental. On such platforms, references to UTF-8 encoding in this
1222 document and elsewhere should be read as meaning the UTF-EBCDIC
1223 specified in Unicode Technical Report 16, unless ASCII vs. EBCDIC issues
1224 are specifically discussed. There is no C<utfebcdic> pragma or
1225 ":utfebcdic" layer; rather, "utf8" and ":utf8" are reused to mean
1226 the platform's "natural" 8-bit encoding of Unicode. See L<perlebcdic>
1227 for more discussion of the issues.
1231 Usually locale settings and Unicode do not affect each other, but
1232 there are a couple of exceptions:
1238 You can enable automatic UTF-8-ification of your standard file
1239 handles, default C<open()> layer, and C<@ARGV> by using either
1240 the C<-C> command line switch or the C<PERL_UNICODE> environment
1241 variable, see L<perlrun> for the documentation of the C<-C> switch.
1245 Perl tries really hard to work both with Unicode and the old
1246 byte-oriented world. Most often this is nice, but sometimes Perl's
1247 straddling of the proverbial fence causes problems.
1251 =head2 When Unicode Does Not Happen
1253 While Perl does have extensive ways to input and output in Unicode,
1254 and few other 'entry points' like the @ARGV which can be interpreted
1255 as Unicode (UTF-8), there still are many places where Unicode (in some
1256 encoding or another) could be given as arguments or received as
1257 results, or both, but it is not.
1259 The following are such interfaces. For all of these interfaces Perl
1260 currently (as of 5.8.3) simply assumes byte strings both as arguments
1261 and results, or UTF-8 strings if the C<encoding> pragma has been used.
1263 One reason why Perl does not attempt to resolve the role of Unicode in
1264 this cases is that the answers are highly dependent on the operating
1265 system and the file system(s). For example, whether filenames can be
1266 in Unicode, and in exactly what kind of encoding, is not exactly a
1267 portable concept. Similarly for the qx and system: how well will the
1268 'command line interface' (and which of them?) handle Unicode?
1274 chdir, chmod, chown, chroot, exec, link, lstat, mkdir,
1275 rename, rmdir, stat, symlink, truncate, unlink, utime, -X
1287 open, opendir, sysopen
1291 qx (aka the backtick operator), system
1299 =head2 Forcing Unicode in Perl (Or Unforcing Unicode in Perl)
1301 Sometimes (see L</"When Unicode Does Not Happen">) there are
1302 situations where you simply need to force Perl to believe that a byte
1303 string is UTF-8, or vice versa. The low-level calls
1304 utf8::upgrade($bytestring) and utf8::downgrade($utf8string) are
1307 Do not use them without careful thought, though: Perl may easily get
1308 very confused, angry, or even crash, if you suddenly change the 'nature'
1309 of scalar like that. Especially careful you have to be if you use the
1310 utf8::upgrade(): any random byte string is not valid UTF-8.
1312 =head2 Using Unicode in XS
1314 If you want to handle Perl Unicode in XS extensions, you may find the
1315 following C APIs useful. See also L<perlguts/"Unicode Support"> for an
1316 explanation about Unicode at the XS level, and L<perlapi> for the API
1323 C<DO_UTF8(sv)> returns true if the C<UTF8> flag is on and the bytes
1324 pragma is not in effect. C<SvUTF8(sv)> returns true is the C<UTF8>
1325 flag is on; the bytes pragma is ignored. The C<UTF8> flag being on
1326 does B<not> mean that there are any characters of code points greater
1327 than 255 (or 127) in the scalar or that there are even any characters
1328 in the scalar. What the C<UTF8> flag means is that the sequence of
1329 octets in the representation of the scalar is the sequence of UTF-8
1330 encoded code points of the characters of a string. The C<UTF8> flag
1331 being off means that each octet in this representation encodes a
1332 single character with code point 0..255 within the string. Perl's
1333 Unicode model is not to use UTF-8 until it is absolutely necessary.
1337 C<uvuni_to_utf8(buf, chr)> writes a Unicode character code point into
1338 a buffer encoding the code point as UTF-8, and returns a pointer
1339 pointing after the UTF-8 bytes.
1343 C<utf8_to_uvuni(buf, lenp)> reads UTF-8 encoded bytes from a buffer and
1344 returns the Unicode character code point and, optionally, the length of
1345 the UTF-8 byte sequence.
1349 C<utf8_length(start, end)> returns the length of the UTF-8 encoded buffer
1350 in characters. C<sv_len_utf8(sv)> returns the length of the UTF-8 encoded
1355 C<sv_utf8_upgrade(sv)> converts the string of the scalar to its UTF-8
1356 encoded form. C<sv_utf8_downgrade(sv)> does the opposite, if
1357 possible. C<sv_utf8_encode(sv)> is like sv_utf8_upgrade except that
1358 it does not set the C<UTF8> flag. C<sv_utf8_decode()> does the
1359 opposite of C<sv_utf8_encode()>. Note that none of these are to be
1360 used as general-purpose encoding or decoding interfaces: C<use Encode>
1361 for that. C<sv_utf8_upgrade()> is affected by the encoding pragma
1362 but C<sv_utf8_downgrade()> is not (since the encoding pragma is
1363 designed to be a one-way street).
1367 C<is_utf8_char(s)> returns true if the pointer points to a valid UTF-8
1372 C<is_utf8_string(buf, len)> returns true if C<len> bytes of the buffer
1377 C<UTF8SKIP(buf)> will return the number of bytes in the UTF-8 encoded
1378 character in the buffer. C<UNISKIP(chr)> will return the number of bytes
1379 required to UTF-8-encode the Unicode character code point. C<UTF8SKIP()>
1380 is useful for example for iterating over the characters of a UTF-8
1381 encoded buffer; C<UNISKIP()> is useful, for example, in computing
1382 the size required for a UTF-8 encoded buffer.
1386 C<utf8_distance(a, b)> will tell the distance in characters between the
1387 two pointers pointing to the same UTF-8 encoded buffer.
1391 C<utf8_hop(s, off)> will return a pointer to an UTF-8 encoded buffer
1392 that is C<off> (positive or negative) Unicode characters displaced
1393 from the UTF-8 buffer C<s>. Be careful not to overstep the buffer:
1394 C<utf8_hop()> will merrily run off the end or the beginning of the
1395 buffer if told to do so.
1399 C<pv_uni_display(dsv, spv, len, pvlim, flags)> and
1400 C<sv_uni_display(dsv, ssv, pvlim, flags)> are useful for debugging the
1401 output of Unicode strings and scalars. By default they are useful
1402 only for debugging--they display B<all> characters as hexadecimal code
1403 points--but with the flags C<UNI_DISPLAY_ISPRINT>,
1404 C<UNI_DISPLAY_BACKSLASH>, and C<UNI_DISPLAY_QQ> you can make the
1405 output more readable.
1409 C<ibcmp_utf8(s1, pe1, u1, l1, u1, s2, pe2, l2, u2)> can be used to
1410 compare two strings case-insensitively in Unicode. For case-sensitive
1411 comparisons you can just use C<memEQ()> and C<memNE()> as usual.
1415 For more information, see L<perlapi>, and F<utf8.c> and F<utf8.h>
1416 in the Perl source code distribution.
1420 =head2 Interaction with Locales
1422 Use of locales with Unicode data may lead to odd results. Currently,
1423 Perl attempts to attach 8-bit locale info to characters in the range
1424 0..255, but this technique is demonstrably incorrect for locales that
1425 use characters above that range when mapped into Unicode. Perl's
1426 Unicode support will also tend to run slower. Use of locales with
1427 Unicode is discouraged.
1429 =head2 Interaction with Extensions
1431 When Perl exchanges data with an extension, the extension should be
1432 able to understand the UTF8 flag and act accordingly. If the
1433 extension doesn't know about the flag, it's likely that the extension
1434 will return incorrectly-flagged data.
1436 So if you're working with Unicode data, consult the documentation of
1437 every module you're using if there are any issues with Unicode data
1438 exchange. If the documentation does not talk about Unicode at all,
1439 suspect the worst and probably look at the source to learn how the
1440 module is implemented. Modules written completely in Perl shouldn't
1441 cause problems. Modules that directly or indirectly access code written
1442 in other programming languages are at risk.
1444 For affected functions, the simple strategy to avoid data corruption is
1445 to always make the encoding of the exchanged data explicit. Choose an
1446 encoding that you know the extension can handle. Convert arguments passed
1447 to the extensions to that encoding and convert results back from that
1448 encoding. Write wrapper functions that do the conversions for you, so
1449 you can later change the functions when the extension catches up.
1451 To provide an example, let's say the popular Foo::Bar::escape_html
1452 function doesn't deal with Unicode data yet. The wrapper function
1453 would convert the argument to raw UTF-8 and convert the result back to
1454 Perl's internal representation like so:
1456 sub my_escape_html ($) {
1458 return unless defined $what;
1459 Encode::decode_utf8(Foo::Bar::escape_html(Encode::encode_utf8($what)));
1462 Sometimes, when the extension does not convert data but just stores
1463 and retrieves them, you will be in a position to use the otherwise
1464 dangerous Encode::_utf8_on() function. Let's say the popular
1465 C<Foo::Bar> extension, written in C, provides a C<param> method that
1466 lets you store and retrieve data according to these prototypes:
1468 $self->param($name, $value); # set a scalar
1469 $value = $self->param($name); # retrieve a scalar
1471 If it does not yet provide support for any encoding, one could write a
1472 derived class with such a C<param> method:
1475 my($self,$name,$value) = @_;
1476 utf8::upgrade($name); # make sure it is UTF-8 encoded
1477 if (defined $value) {
1478 utf8::upgrade($value); # make sure it is UTF-8 encoded
1479 return $self->SUPER::param($name,$value);
1481 my $ret = $self->SUPER::param($name);
1482 Encode::_utf8_on($ret); # we know, it is UTF-8 encoded
1487 Some extensions provide filters on data entry/exit points, such as
1488 DB_File::filter_store_key and family. Look out for such filters in
1489 the documentation of your extensions, they can make the transition to
1490 Unicode data much easier.
1494 Some functions are slower when working on UTF-8 encoded strings than
1495 on byte encoded strings. All functions that need to hop over
1496 characters such as length(), substr() or index(), or matching regular
1497 expressions can work B<much> faster when the underlying data are
1500 In Perl 5.8.0 the slowness was often quite spectacular; in Perl 5.8.1
1501 a caching scheme was introduced which will hopefully make the slowness
1502 somewhat less spectacular, at least for some operations. In general,
1503 operations with UTF-8 encoded strings are still slower. As an example,
1504 the Unicode properties (character classes) like C<\p{Nd}> are known to
1505 be quite a bit slower (5-20 times) than their simpler counterparts
1506 like C<\d> (then again, there 268 Unicode characters matching C<Nd>
1507 compared with the 10 ASCII characters matching C<d>).
1509 =head2 Porting code from perl-5.6.X
1511 Perl 5.8 has a different Unicode model from 5.6. In 5.6 the programmer
1512 was required to use the C<utf8> pragma to declare that a given scope
1513 expected to deal with Unicode data and had to make sure that only
1514 Unicode data were reaching that scope. If you have code that is
1515 working with 5.6, you will need some of the following adjustments to
1516 your code. The examples are written such that the code will continue
1517 to work under 5.6, so you should be safe to try them out.
1523 A filehandle that should read or write UTF-8
1526 binmode $fh, ":utf8";
1531 A scalar that is going to be passed to some extension
1533 Be it Compress::Zlib, Apache::Request or any extension that has no
1534 mention of Unicode in the manpage, you need to make sure that the
1535 UTF8 flag is stripped off. Note that at the time of this writing
1536 (October 2002) the mentioned modules are not UTF-8-aware. Please
1537 check the documentation to verify if this is still true.
1541 $val = Encode::encode_utf8($val); # make octets
1546 A scalar we got back from an extension
1548 If you believe the scalar comes back as UTF-8, you will most likely
1549 want the UTF8 flag restored:
1553 $val = Encode::decode_utf8($val);
1558 Same thing, if you are really sure it is UTF-8
1562 Encode::_utf8_on($val);
1567 A wrapper for fetchrow_array and fetchrow_hashref
1569 When the database contains only UTF-8, a wrapper function or method is
1570 a convenient way to replace all your fetchrow_array and
1571 fetchrow_hashref calls. A wrapper function will also make it easier to
1572 adapt to future enhancements in your database driver. Note that at the
1573 time of this writing (October 2002), the DBI has no standardized way
1574 to deal with UTF-8 data. Please check the documentation to verify if
1578 my($self, $sth, $what) = @_; # $what is one of fetchrow_{array,hashref}
1584 my @arr = $sth->$what;
1586 defined && /[^\000-\177]/ && Encode::_utf8_on($_);
1590 my $ret = $sth->$what;
1592 for my $k (keys %$ret) {
1593 defined && /[^\000-\177]/ && Encode::_utf8_on($_) for $ret->{$k};
1597 defined && /[^\000-\177]/ && Encode::_utf8_on($_) for $ret;
1607 A large scalar that you know can only contain ASCII
1609 Scalars that contain only ASCII and are marked as UTF-8 are sometimes
1610 a drag to your program. If you recognize such a situation, just remove
1613 utf8::downgrade($val) if $] > 5.007;
1619 L<perlunitut>, L<perluniintro>, L<Encode>, L<open>, L<utf8>, L<bytes>,
1620 L<perlretut>, L<perlvar/"${^UNICODE}">