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.
15 =item Input and Output Layers
17 Perl knows when a filehandle uses Perl's internal Unicode encodings
18 (UTF-8, or UTF-EBCDIC if in EBCDIC) if the filehandle is opened with
19 the ":utf8" layer. Other encodings can be converted to Perl's
20 encoding on input or from Perl's encoding on output by use of the
21 ":encoding(...)" layer. See L<open>.
23 To indicate that Perl source itself is using a particular encoding,
26 =item Regular Expressions
28 The regular expression compiler produces polymorphic opcodes. That is,
29 the pattern adapts to the data and automatically switches to the Unicode
30 character scheme when presented with Unicode data--or instead uses
31 a traditional byte scheme when presented with byte data.
33 =item C<use utf8> still needed to enable UTF-8/UTF-EBCDIC in scripts
35 As a compatibility measure, the C<use utf8> pragma must be explicitly
36 included to enable recognition of UTF-8 in the Perl scripts themselves
37 (in string or regular expression literals, or in identifier names) on
38 ASCII-based machines or to recognize UTF-EBCDIC on EBCDIC-based
39 machines. B<These are the only times when an explicit C<use utf8>
40 is needed.> See L<utf8>.
42 You can also use the C<encoding> pragma to change the default encoding
43 of the data in your script; see L<encoding>.
45 =item C<use encoding> needed to upgrade non-Latin-1 byte strings
47 By default, there is a fundamental asymmetry in Perl's unicode model:
48 implicit upgrading from byte strings to Unicode strings assumes that
49 they were encoded in I<ISO 8859-1 (Latin-1)>, but Unicode strings are
50 downgraded with UTF-8 encoding. This happens because the first 256
51 codepoints in Unicode happens to agree with Latin-1.
53 If you wish to interpret byte strings as UTF-8 instead, use the
58 See L</"Byte and Character Semantics"> for more details.
62 =head2 Byte and Character Semantics
64 Beginning with version 5.6, Perl uses logically-wide characters to
65 represent strings internally.
67 In future, Perl-level operations will be expected to work with
68 characters rather than bytes.
70 However, as an interim compatibility measure, Perl aims to
71 provide a safe migration path from byte semantics to character
72 semantics for programs. For operations where Perl can unambiguously
73 decide that the input data are characters, Perl switches to
74 character semantics. For operations where this determination cannot
75 be made without additional information from the user, Perl decides in
76 favor of compatibility and chooses to use byte semantics.
78 This behavior preserves compatibility with earlier versions of Perl,
79 which allowed byte semantics in Perl operations only if
80 none of the program's inputs were marked as being as source of Unicode
81 character data. Such data may come from filehandles, from calls to
82 external programs, from information provided by the system (such as %ENV),
83 or from literals and constants in the source text.
85 The C<bytes> pragma will always, regardless of platform, force byte
86 semantics in a particular lexical scope. See L<bytes>.
88 The C<utf8> pragma is primarily a compatibility device that enables
89 recognition of UTF-(8|EBCDIC) in literals encountered by the parser.
90 Note that this pragma is only required while Perl defaults to byte
91 semantics; when character semantics become the default, this pragma
92 may become a no-op. See L<utf8>.
94 Unless explicitly stated, Perl operators use character semantics
95 for Unicode data and byte semantics for non-Unicode data.
96 The decision to use character semantics is made transparently. If
97 input data comes from a Unicode source--for example, if a character
98 encoding layer is added to a filehandle or a literal Unicode
99 string constant appears in a program--character semantics apply.
100 Otherwise, byte semantics are in effect. The C<bytes> pragma should
101 be used to force byte semantics on Unicode data.
103 If strings operating under byte semantics and strings with Unicode
104 character data are concatenated, the new string will be created by
105 decoding the byte strings as I<ISO 8859-1 (Latin-1)>, even if the
106 old Unicode string used EBCDIC. This translation is done without
107 regard to the system's native 8-bit encoding. To change this for
108 systems with non-Latin-1 and non-EBCDIC native encodings, use the
109 C<encoding> pragma. See L<encoding>.
111 Under character semantics, many operations that formerly operated on
112 bytes now operate on characters. A character in Perl is
113 logically just a number ranging from 0 to 2**31 or so. Larger
114 characters may encode into longer sequences of bytes internally, but
115 this internal detail is mostly hidden for Perl code.
116 See L<perluniintro> for more.
118 =head2 Effects of Character Semantics
120 Character semantics have the following effects:
126 Strings--including hash keys--and regular expression patterns may
127 contain characters that have an ordinal value larger than 255.
129 If you use a Unicode editor to edit your program, Unicode characters
130 may occur directly within the literal strings in one of the various
131 Unicode encodings (UTF-8, UTF-EBCDIC, UCS-2, etc.), but will be recognized
132 as such and converted to Perl's internal representation only if the
133 appropriate L<encoding> is specified.
135 Unicode characters can also be added to a string by using the
136 C<\x{...}> notation. The Unicode code for the desired character, in
137 hexadecimal, should be placed in the braces. For instance, a smiley
138 face is C<\x{263A}>. This encoding scheme only works for characters
139 with a code of 0x100 or above.
143 use charnames ':full';
145 you can use the C<\N{...}> notation and put the official Unicode
146 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. The C<\C> pattern is provided to force
160 a match a single byte--a C<char> in C, hence C<\C>.
164 Character classes in regular expressions match characters instead of
165 bytes and match against the character properties specified in the
166 Unicode properties database. C<\w> can be used to match a Japanese
167 ideograph, for instance.
171 Named Unicode properties, scripts, and block ranges may be used like
172 character classes via the C<\p{}> "matches property" construct and
173 the C<\P{}> negation, "doesn't match property".
175 For instance, C<\p{Lu}> matches any character with the Unicode "Lu"
176 (Letter, uppercase) property, while C<\p{M}> matches any character
177 with an "M" (mark--accents and such) property. Brackets are not
178 required for single letter properties, so C<\p{M}> is equivalent to
179 C<\pM>. Many predefined properties are available, such as
180 C<\p{Mirrored}> and C<\p{Tibetan}>.
182 The official Unicode script and block names have spaces and dashes as
183 separators, but for convenience you can use dashes, spaces, or
184 underbars, and case is unimportant. It is recommended, however, that
185 for consistency you use the following naming: the official Unicode
186 script, property, or block name (see below for the additional rules
187 that apply to block names) with whitespace and dashes removed, and the
188 words "uppercase-first-lowercase-rest". C<Latin-1 Supplement> thus
189 becomes C<Latin1Supplement>.
191 You can also use negation in both C<\p{}> and C<\P{}> by introducing a caret
192 (^) between the first brace and the property name: C<\p{^Tamil}> is
193 equal to C<\P{Tamil}>.
195 B<NOTE: the properties, scripts, and blocks listed here are as of
196 Unicode 3.2.0, March 2002, or Perl 5.8.0, July 2002. Unicode 4.0.0
197 came out in April 2003, and Perl 5.8.1 in September 2003.>
199 Here are the basic Unicode General Category properties, followed by their
200 long form. You can use either; C<\p{Lu}> and C<\p{UppercaseLetter}>,
201 for instance, are identical.
223 Pc ConnectorPunctuation
227 Pi InitialPunctuation
228 (may behave like Ps or Pe depending on usage)
230 (may behave like Ps or Pe depending on usage)
242 Zp ParagraphSeparator
247 Cs Surrogate (not usable)
251 Single-letter properties match all characters in any of the
252 two-letter sub-properties starting with the same letter.
253 C<L&> is a special case, which is an alias for C<Ll>, C<Lu>, and C<Lt>.
255 Because Perl hides the need for the user to understand the internal
256 representation of Unicode characters, there is no need to implement
257 the somewhat messy concept of surrogates. C<Cs> is therefore not
260 Because scripts differ in their directionality--Hebrew is
261 written right to left, for example--Unicode supplies these properties:
266 BidiLRE Left-to-Right Embedding
267 BidiLRO Left-to-Right Override
269 BidiAL Right-to-Left Arabic
270 BidiRLE Right-to-Left Embedding
271 BidiRLO Right-to-Left Override
272 BidiPDF Pop Directional Format
273 BidiEN European Number
274 BidiES European Number Separator
275 BidiET European Number Terminator
277 BidiCS Common Number Separator
278 BidiNSM Non-Spacing Mark
279 BidiBN Boundary Neutral
280 BidiB Paragraph Separator
281 BidiS Segment Separator
283 BidiON Other Neutrals
285 For example, C<\p{BidiR}> matches characters that are normally
286 written right to left.
292 The script names which can be used by C<\p{...}> and C<\P{...}>,
293 such as in C<\p{Latin}> or C<\p{Cyrillic}>, are as follows:
340 Extended property classes can supplement the basic
341 properties, defined by the F<PropList> Unicode database:
356 LogicalOrderException
357 NoncharacterCodePoint
359 OtherDefaultIgnorableCodePoint
371 and there are further derived properties:
373 Alphabetic Lu + Ll + Lt + Lm + Lo + OtherAlphabetic
374 Lowercase Ll + OtherLowercase
375 Uppercase Lu + OtherUppercase
378 ID_Start Lu + Ll + Lt + Lm + Lo + Nl
379 ID_Continue ID_Start + Mn + Mc + Nd + Pc
382 Assigned Any non-Cn character (i.e. synonym for \P{Cn})
383 Unassigned Synonym for \p{Cn}
384 Common Any character (or unassigned code point)
385 not explicitly assigned to a script
387 For backward compatibility (with Perl 5.6), all properties mentioned
388 so far may have C<Is> prepended to their name, so C<\P{IsLu}>, for
389 example, is equal to C<\P{Lu}>.
393 In addition to B<scripts>, Unicode also defines B<blocks> of
394 characters. The difference between scripts and blocks is that the
395 concept of scripts is closer to natural languages, while the concept
396 of blocks is more of an artificial grouping based on groups of 256
397 Unicode characters. For example, the C<Latin> script contains letters
398 from many blocks but does not contain all the characters from those
399 blocks. It does not, for example, contain digits, because digits are
400 shared across many scripts. Digits and similar groups, like
401 punctuation, are in a category called C<Common>.
403 For more about scripts, see the UTR #24:
405 http://www.unicode.org/unicode/reports/tr24/
407 For more about blocks, see:
409 http://www.unicode.org/Public/UNIDATA/Blocks.txt
411 Block names are given with the C<In> prefix. For example, the
412 Katakana block is referenced via C<\p{InKatakana}>. The C<In>
413 prefix may be omitted if there is no naming conflict with a script
414 or any other property, but it is recommended that C<In> always be used
415 for block tests to avoid confusion.
417 These block names are supported:
419 InAlphabeticPresentationForms
421 InArabicPresentationFormsA
422 InArabicPresentationFormsB
433 InByzantineMusicalSymbols
435 InCJKCompatibilityForms
436 InCJKCompatibilityIdeographs
437 InCJKCompatibilityIdeographsSupplement
438 InCJKRadicalsSupplement
439 InCJKSymbolsAndPunctuation
440 InCJKUnifiedIdeographs
441 InCJKUnifiedIdeographsExtensionA
442 InCJKUnifiedIdeographsExtensionB
444 InCombiningDiacriticalMarks
445 InCombiningDiacriticalMarksforSymbols
450 InCyrillicSupplementary
454 InEnclosedAlphanumerics
455 InEnclosedCJKLettersAndMonths
465 InHalfwidthAndFullwidthForms
466 InHangulCompatibilityJamo
471 InHighPrivateUseSurrogates
475 InIdeographicDescriptionCharacters
480 InKatakanaPhoneticExtensions
485 InLatinExtendedAdditional
490 InMathematicalAlphanumericSymbols
491 InMathematicalOperators
492 InMiscellaneousMathematicalSymbolsA
493 InMiscellaneousMathematicalSymbolsB
494 InMiscellaneousSymbols
495 InMiscellaneousTechnical
502 InOpticalCharacterRecognition
508 InSpacingModifierLetters
510 InSuperscriptsAndSubscripts
511 InSupplementalArrowsA
512 InSupplementalArrowsB
513 InSupplementalMathematicalOperators
514 InSupplementaryPrivateUseAreaA
515 InSupplementaryPrivateUseAreaB
525 InUnifiedCanadianAboriginalSyllabics
534 The special pattern C<\X> matches any extended Unicode
535 sequence--"a combining character sequence" in Standardese--where the
536 first character is a base character and subsequent characters are mark
537 characters that apply to the base character. C<\X> is equivalent to
542 The C<tr///> operator translates characters instead of bytes. Note
543 that the C<tr///CU> functionality has been removed. For similar
544 functionality see pack('U0', ...) and pack('C0', ...).
548 Case translation operators use the Unicode case translation tables
549 when character input is provided. Note that C<uc()>, or C<\U> in
550 interpolated strings, translates to uppercase, while C<ucfirst>,
551 or C<\u> in interpolated strings, translates to titlecase in languages
552 that make the distinction.
556 Most operators that deal with positions or lengths in a string will
557 automatically switch to using character positions, including
558 C<chop()>, C<substr()>, C<pos()>, C<index()>, C<rindex()>,
559 C<sprintf()>, C<write()>, and C<length()>. Operators that
560 specifically do not switch include C<vec()>, C<pack()>, and
561 C<unpack()>. Operators that really don't care include C<chomp()>,
562 operators that treats strings as a bucket of bits such as C<sort()>,
563 and operators dealing with filenames.
567 The C<pack()>/C<unpack()> letters C<c> and C<C> do I<not> change,
568 since they are often used for byte-oriented formats. Again, think
569 C<char> in the C language.
571 There is a new C<U> specifier that converts between Unicode characters
576 The C<chr()> and C<ord()> functions work on characters, similar to
577 C<pack("U")> and C<unpack("U")>, I<not> C<pack("C")> and
578 C<unpack("C")>. C<pack("C")> and C<unpack("C")> are methods for
579 emulating byte-oriented C<chr()> and C<ord()> on Unicode strings.
580 While these methods reveal the internal encoding of Unicode strings,
581 that is not something one normally needs to care about at all.
585 The bit string operators, C<& | ^ ~>, can operate on character data.
586 However, for backward compatibility, such as when using bit string
587 operations when characters are all less than 256 in ordinal value, one
588 should not use C<~> (the bit complement) with characters of both
589 values less than 256 and values greater than 256. Most importantly,
590 DeMorgan's laws (C<~($x|$y) eq ~$x&~$y> and C<~($x&$y) eq ~$x|~$y>)
591 will not hold. The reason for this mathematical I<faux pas> is that
592 the complement cannot return B<both> the 8-bit (byte-wide) bit
593 complement B<and> the full character-wide bit complement.
597 lc(), uc(), lcfirst(), and ucfirst() work for the following cases:
603 the case mapping is from a single Unicode character to another
604 single Unicode character, or
608 the case mapping is from a single Unicode character to more
609 than one Unicode character.
613 Things to do with locales (Lithuanian, Turkish, Azeri) do B<not> work
614 since Perl does not understand the concept of Unicode locales.
616 See the Unicode Technical Report #21, Case Mappings, for more details.
624 And finally, C<scalar reverse()> reverses by character rather than by byte.
628 =head2 User-Defined Character Properties
630 You can define your own character properties by defining subroutines
631 whose names begin with "In" or "Is". The subroutines must be defined
632 in the C<main> package. The user-defined properties can be used in the
633 regular expression C<\p> and C<\P> constructs. Note that the effect
634 is compile-time and immutable once defined.
636 The subroutines must return a specially-formatted string, with one
637 or more newline-separated lines. Each line must be one of the following:
643 Two hexadecimal numbers separated by horizontal whitespace (space or
644 tabular characters) denoting a range of Unicode code points to include.
648 Something to include, prefixed by "+": a built-in character
649 property (prefixed by "utf8::"), to represent all the characters in that
650 property; two hexadecimal code points for a range; or a single
651 hexadecimal code point.
655 Something to exclude, prefixed by "-": an existing character
656 property (prefixed by "utf8::"), for all the characters in that
657 property; two hexadecimal code points for a range; or a single
658 hexadecimal code point.
662 Something to negate, prefixed "!": an existing character
663 property (prefixed by "utf8::") for all the characters except the
664 characters in the property; two hexadecimal code points for a range;
665 or a single hexadecimal code point.
669 For example, to define a property that covers both the Japanese
670 syllabaries (hiragana and katakana), you can define
679 Imagine that the here-doc end marker is at the beginning of the line.
680 Now you can use C<\p{InKana}> and C<\P{InKana}>.
682 You could also have used the existing block property names:
691 Suppose you wanted to match only the allocated characters,
692 not the raw block ranges: in other words, you want to remove
703 The negation is useful for defining (surprise!) negated classes.
713 You can also define your own mappings to be used in the lc(),
714 lcfirst(), uc(), and ucfirst() (or their string-inlined versions).
715 The principle is the same: define subroutines in the C<main> package
716 with names like C<ToLower> (for lc() and lcfirst()), C<ToTitle> (for
717 the first character in ucfirst()), and C<ToUpper> (for uc(), and the
718 rest of the characters in ucfirst()).
720 The string returned by the subroutines needs now to be three
721 hexadecimal numbers separated by tabulators: start of the source
722 range, end of the source range, and start of the destination range.
731 defines an uc() mapping that causes only the characters "a", "b", and
732 "c" to be mapped to "A", "B", "C", all other characters will remain
735 If there is no source range to speak of, that is, the mapping is from
736 a single character to another single character, leave the end of the
737 source range empty, but the two tabulator characters are still needed.
746 defines a lc() mapping that causes only "A" to be mapped to "a", all
747 other characters will remain unchanged.
749 (For serious hackers only) If you want to introspect the default
750 mappings, you can find the data in the directory
751 C<$Config{privlib}>/F<unicore/To/>. The mapping data is returned as
752 the here-document, and the C<utf8::ToSpecFoo> are special exception
753 mappings derived from <$Config{privlib}>/F<unicore/SpecialCasing.txt>.
754 The C<Digit> and C<Fold> mappings that one can see in the directory
755 are not directly user-accessible, one can use either the
756 C<Unicode::UCD> module, or just match case-insensitively (that's when
757 the C<Fold> mapping is used).
759 A final note on the user-defined property tests and mappings: they
760 will be used only if the scalar has been marked as having Unicode
761 characters. Old byte-style strings will not be affected.
763 =head2 Character Encodings for Input and Output
767 =head2 Unicode Regular Expression Support Level
769 The following list of Unicode support for regular expressions describes
770 all the features currently supported. The references to "Level N"
771 and the section numbers refer to the Unicode Technical Report 18,
772 "Unicode Regular Expression Guidelines", version 6 (Unicode 3.2.0,
779 Level 1 - Basic Unicode Support
781 2.1 Hex Notation - done [1]
782 Named Notation - done [2]
783 2.2 Categories - done [3][4]
784 2.3 Subtraction - MISSING [5][6]
785 2.4 Simple Word Boundaries - done [7]
786 2.5 Simple Loose Matches - done [8]
787 2.6 End of Line - MISSING [9][10]
791 [ 3] . \p{...} \P{...}
792 [ 4] now scripts (see UTR#24 Script Names) in addition to blocks
794 [ 6] can use regular expression look-ahead [a]
795 or user-defined character properties [b] to emulate subtraction
796 [ 7] include Letters in word characters
797 [ 8] note that Perl does Full case-folding in matching, not Simple:
798 for example U+1F88 is equivalent with U+1F00 U+03B9,
799 not with 1F80. This difference matters for certain Greek
800 capital letters with certain modifiers: the Full case-folding
801 decomposes the letter, while the Simple case-folding would map
802 it to a single character.
803 [ 9] see UTR #13 Unicode Newline Guidelines
804 [10] should do ^ and $ also on \x{85}, \x{2028} and \x{2029}
805 (should also affect <>, $., and script line numbers)
806 (the \x{85}, \x{2028} and \x{2029} do match \s)
808 [a] You can mimic class subtraction using lookahead.
809 For example, what UTR #18 might write as
811 [{Greek}-[{UNASSIGNED}]]
813 in Perl can be written as:
815 (?!\p{Unassigned})\p{InGreekAndCoptic}
816 (?=\p{Assigned})\p{InGreekAndCoptic}
818 But in this particular example, you probably really want
822 which will match assigned characters known to be part of the Greek script.
824 Also see the Unicode::Regex::Set module, it does implement the full
825 UTR #18 grouping, intersection, union, and removal (subtraction) syntax.
827 [b] See L</"User-Defined Character Properties">.
831 Level 2 - Extended Unicode Support
833 3.1 Surrogates - MISSING [11]
834 3.2 Canonical Equivalents - MISSING [12][13]
835 3.3 Locale-Independent Graphemes - MISSING [14]
836 3.4 Locale-Independent Words - MISSING [15]
837 3.5 Locale-Independent Loose Matches - MISSING [16]
839 [11] Surrogates are solely a UTF-16 concept and Perl's internal
840 representation is UTF-8. The Encode module does UTF-16, though.
841 [12] see UTR#15 Unicode Normalization
842 [13] have Unicode::Normalize but not integrated to regexes
843 [14] have \X but at this level . should equal that
844 [15] need three classes, not just \w and \W
845 [16] see UTR#21 Case Mappings
849 Level 3 - Locale-Sensitive Support
851 4.1 Locale-Dependent Categories - MISSING
852 4.2 Locale-Dependent Graphemes - MISSING [16][17]
853 4.3 Locale-Dependent Words - MISSING
854 4.4 Locale-Dependent Loose Matches - MISSING
855 4.5 Locale-Dependent Ranges - MISSING
857 [16] see UTR#10 Unicode Collation Algorithms
858 [17] have Unicode::Collate but not integrated to regexes
862 =head2 Unicode Encodings
864 Unicode characters are assigned to I<code points>, which are abstract
865 numbers. To use these numbers, various encodings are needed.
873 UTF-8 is a variable-length (1 to 6 bytes, current character allocations
874 require 4 bytes), byte-order independent encoding. For ASCII (and we
875 really do mean 7-bit ASCII, not another 8-bit encoding), UTF-8 is
878 The following table is from Unicode 3.2.
880 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
882 U+0000..U+007F 00..7F
883 U+0080..U+07FF C2..DF 80..BF
884 U+0800..U+0FFF E0 A0..BF 80..BF
885 U+1000..U+CFFF E1..EC 80..BF 80..BF
886 U+D000..U+D7FF ED 80..9F 80..BF
887 U+D800..U+DFFF ******* ill-formed *******
888 U+E000..U+FFFF EE..EF 80..BF 80..BF
889 U+10000..U+3FFFF F0 90..BF 80..BF 80..BF
890 U+40000..U+FFFFF F1..F3 80..BF 80..BF 80..BF
891 U+100000..U+10FFFF F4 80..8F 80..BF 80..BF
893 Note the C<A0..BF> in C<U+0800..U+0FFF>, the C<80..9F> in
894 C<U+D000...U+D7FF>, the C<90..B>F in C<U+10000..U+3FFFF>, and the
895 C<80...8F> in C<U+100000..U+10FFFF>. The "gaps" are caused by legal
896 UTF-8 avoiding non-shortest encodings: it is technically possible to
897 UTF-8-encode a single code point in different ways, but that is
898 explicitly forbidden, and the shortest possible encoding should always
899 be used. So that's what Perl does.
901 Another way to look at it is via bits:
903 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
906 00000bbbbbaaaaaa 110bbbbb 10aaaaaa
907 ccccbbbbbbaaaaaa 1110cccc 10bbbbbb 10aaaaaa
908 00000dddccccccbbbbbbaaaaaa 11110ddd 10cccccc 10bbbbbb 10aaaaaa
910 As you can see, the continuation bytes all begin with C<10>, and the
911 leading bits of the start byte tell how many bytes the are in the
918 Like UTF-8 but EBCDIC-safe, in the way that UTF-8 is ASCII-safe.
922 UTF-16, UTF-16BE, UTF-16LE, Surrogates, and BOMs (Byte Order Marks)
924 The followings items are mostly for reference and general Unicode
925 knowledge, Perl doesn't use these constructs internally.
927 UTF-16 is a 2 or 4 byte encoding. The Unicode code points
928 C<U+0000..U+FFFF> are stored in a single 16-bit unit, and the code
929 points C<U+10000..U+10FFFF> in two 16-bit units. The latter case is
930 using I<surrogates>, the first 16-bit unit being the I<high
931 surrogate>, and the second being the I<low surrogate>.
933 Surrogates are code points set aside to encode the C<U+10000..U+10FFFF>
934 range of Unicode code points in pairs of 16-bit units. The I<high
935 surrogates> are the range C<U+D800..U+DBFF>, and the I<low surrogates>
936 are the range C<U+DC00..U+DFFF>. The surrogate encoding is
938 $hi = ($uni - 0x10000) / 0x400 + 0xD800;
939 $lo = ($uni - 0x10000) % 0x400 + 0xDC00;
943 $uni = 0x10000 + ($hi - 0xD800) * 0x400 + ($lo - 0xDC00);
945 If you try to generate surrogates (for example by using chr()), you
946 will get a warning if warnings are turned on, because those code
947 points are not valid for a Unicode character.
949 Because of the 16-bitness, UTF-16 is byte-order dependent. UTF-16
950 itself can be used for in-memory computations, but if storage or
951 transfer is required either UTF-16BE (big-endian) or UTF-16LE
952 (little-endian) encodings must be chosen.
954 This introduces another problem: what if you just know that your data
955 is UTF-16, but you don't know which endianness? Byte Order Marks, or
956 BOMs, are a solution to this. A special character has been reserved
957 in Unicode to function as a byte order marker: the character with the
958 code point C<U+FEFF> is the BOM.
960 The trick is that if you read a BOM, you will know the byte order,
961 since if it was written on a big-endian platform, you will read the
962 bytes C<0xFE 0xFF>, but if it was written on a little-endian platform,
963 you will read the bytes C<0xFF 0xFE>. (And if the originating platform
964 was writing in UTF-8, you will read the bytes C<0xEF 0xBB 0xBF>.)
966 The way this trick works is that the character with the code point
967 C<U+FFFE> is guaranteed not to be a valid Unicode character, so the
968 sequence of bytes C<0xFF 0xFE> is unambiguously "BOM, represented in
969 little-endian format" and cannot be C<U+FFFE>, represented in big-endian
974 UTF-32, UTF-32BE, UTF-32LE
976 The UTF-32 family is pretty much like the UTF-16 family, expect that
977 the units are 32-bit, and therefore the surrogate scheme is not
978 needed. The BOM signatures will be C<0x00 0x00 0xFE 0xFF> for BE and
979 C<0xFF 0xFE 0x00 0x00> for LE.
985 Encodings defined by the ISO 10646 standard. UCS-2 is a 16-bit
986 encoding. Unlike UTF-16, UCS-2 is not extensible beyond C<U+FFFF>,
987 because it does not use surrogates. UCS-4 is a 32-bit encoding,
988 functionally identical to UTF-32.
994 A seven-bit safe (non-eight-bit) encoding, which is useful if the
995 transport or storage is not eight-bit safe. Defined by RFC 2152.
999 =head2 Security Implications of Unicode
1007 Unfortunately, the specification of UTF-8 leaves some room for
1008 interpretation of how many bytes of encoded output one should generate
1009 from one input Unicode character. Strictly speaking, the shortest
1010 possible sequence of UTF-8 bytes should be generated,
1011 because otherwise there is potential for an input buffer overflow at
1012 the receiving end of a UTF-8 connection. Perl always generates the
1013 shortest length UTF-8, and with warnings on Perl will warn about
1014 non-shortest length UTF-8 along with other malformations, such as the
1015 surrogates, which are not real Unicode code points.
1019 Regular expressions behave slightly differently between byte data and
1020 character (Unicode) data. For example, the "word character" character
1021 class C<\w> will work differently depending on if data is eight-bit bytes
1024 In the first case, the set of C<\w> characters is either small--the
1025 default set of alphabetic characters, digits, and the "_"--or, if you
1026 are using a locale (see L<perllocale>), the C<\w> might contain a few
1027 more letters according to your language and country.
1029 In the second case, the C<\w> set of characters is much, much larger.
1030 Most importantly, even in the set of the first 256 characters, it will
1031 probably match different characters: unlike most locales, which are
1032 specific to a language and country pair, Unicode classifies all the
1033 characters that are letters I<somewhere> as C<\w>. For example, your
1034 locale might not think that LATIN SMALL LETTER ETH is a letter (unless
1035 you happen to speak Icelandic), but Unicode does.
1037 As discussed elsewhere, Perl has one foot (two hooves?) planted in
1038 each of two worlds: the old world of bytes and the new world of
1039 characters, upgrading from bytes to characters when necessary.
1040 If your legacy code does not explicitly use Unicode, no automatic
1041 switch-over to characters should happen. Characters shouldn't get
1042 downgraded to bytes, either. It is possible to accidentally mix bytes
1043 and characters, however (see L<perluniintro>), in which case C<\w> in
1044 regular expressions might start behaving differently. Review your
1045 code. Use warnings and the C<strict> pragma.
1049 =head2 Unicode in Perl on EBCDIC
1051 The way Unicode is handled on EBCDIC platforms is still
1052 experimental. On such platforms, references to UTF-8 encoding in this
1053 document and elsewhere should be read as meaning the UTF-EBCDIC
1054 specified in Unicode Technical Report 16, unless ASCII vs. EBCDIC issues
1055 are specifically discussed. There is no C<utfebcdic> pragma or
1056 ":utfebcdic" layer; rather, "utf8" and ":utf8" are reused to mean
1057 the platform's "natural" 8-bit encoding of Unicode. See L<perlebcdic>
1058 for more discussion of the issues.
1062 Usually locale settings and Unicode do not affect each other, but
1063 there are a couple of exceptions:
1069 You can enable automatic UTF-8-ification of your standard file
1070 handles, default C<open()> layer, and C<@ARGV> by using either
1071 the C<-C> command line switch or the C<PERL_UNICODE> environment
1072 variable, see L<perlrun> for the documentation of the C<-C> switch.
1076 Perl tries really hard to work both with Unicode and the old
1077 byte-oriented world. Most often this is nice, but sometimes Perl's
1078 straddling of the proverbial fence causes problems.
1082 =head2 When Unicode Does Not Happen
1084 While Perl does have extensive ways to input and output in Unicode,
1085 and few other 'entry points' like the @ARGV which can be interpreted
1086 as Unicode (UTF-8), there still are many places where Unicode (in some
1087 encoding or another) could be given as arguments or received as
1088 results, or both, but it is not.
1090 The following are such interfaces. For all of these Perl currently
1091 (as of 5.8.1) simply assumes byte strings both as arguments and results.
1093 One reason why Perl does not attempt to resolve the role of Unicode in
1094 this cases is that the answers are highly dependent on the operating
1095 system and the file system(s). For example, whether filenames can be
1096 in Unicode, and in exactly what kind of encoding, is not exactly a
1097 portable concept. Similarly for the qx and system: how well will the
1098 'command line interface' (and which of them?) handle Unicode?
1104 chmod, chmod, chown, chroot, exec, link, mkdir
1105 rename, rmdir stat, symlink, truncate, unlink, utime
1117 open, opendir, sysopen
1121 qx (aka the backtick operator), system
1129 =head2 Forcing Unicode in Perl (Or Unforcing Unicode in Perl)
1131 Sometimes (see L</"When Unicode Does Not Happen">) there are
1132 situations where you simply need to force Perl to believe that a byte
1133 string is UTF-8, or vice versa. The low-level calls
1134 utf8::upgrade($bytestring) and utf8::downgrade($utf8string) are
1137 Do not use them without careful thought, though: Perl may easily get
1138 very confused, angry, or even crash, if you suddenly change the 'nature'
1139 of scalar like that. Especially careful you have to be if you use the
1140 utf8::upgrade(): any random byte string is not valid UTF-8.
1142 =head2 Using Unicode in XS
1144 If you want to handle Perl Unicode in XS extensions, you may find the
1145 following C APIs useful. See also L<perlguts/"Unicode Support"> for an
1146 explanation about Unicode at the XS level, and L<perlapi> for the API
1153 C<DO_UTF8(sv)> returns true if the C<UTF8> flag is on and the bytes
1154 pragma is not in effect. C<SvUTF8(sv)> returns true is the C<UTF8>
1155 flag is on; the bytes pragma is ignored. The C<UTF8> flag being on
1156 does B<not> mean that there are any characters of code points greater
1157 than 255 (or 127) in the scalar or that there are even any characters
1158 in the scalar. What the C<UTF8> flag means is that the sequence of
1159 octets in the representation of the scalar is the sequence of UTF-8
1160 encoded code points of the characters of a string. The C<UTF8> flag
1161 being off means that each octet in this representation encodes a
1162 single character with code point 0..255 within the string. Perl's
1163 Unicode model is not to use UTF-8 until it is absolutely necessary.
1167 C<uvuni_to_utf8(buf, chr)> writes a Unicode character code point into
1168 a buffer encoding the code point as UTF-8, and returns a pointer
1169 pointing after the UTF-8 bytes.
1173 C<utf8_to_uvuni(buf, lenp)> reads UTF-8 encoded bytes from a buffer and
1174 returns the Unicode character code point and, optionally, the length of
1175 the UTF-8 byte sequence.
1179 C<utf8_length(start, end)> returns the length of the UTF-8 encoded buffer
1180 in characters. C<sv_len_utf8(sv)> returns the length of the UTF-8 encoded
1185 C<sv_utf8_upgrade(sv)> converts the string of the scalar to its UTF-8
1186 encoded form. C<sv_utf8_downgrade(sv)> does the opposite, if
1187 possible. C<sv_utf8_encode(sv)> is like sv_utf8_upgrade except that
1188 it does not set the C<UTF8> flag. C<sv_utf8_decode()> does the
1189 opposite of C<sv_utf8_encode()>. Note that none of these are to be
1190 used as general-purpose encoding or decoding interfaces: C<use Encode>
1191 for that. C<sv_utf8_upgrade()> is affected by the encoding pragma
1192 but C<sv_utf8_downgrade()> is not (since the encoding pragma is
1193 designed to be a one-way street).
1197 C<is_utf8_char(s)> returns true if the pointer points to a valid UTF-8
1202 C<is_utf8_string(buf, len)> returns true if C<len> bytes of the buffer
1207 C<UTF8SKIP(buf)> will return the number of bytes in the UTF-8 encoded
1208 character in the buffer. C<UNISKIP(chr)> will return the number of bytes
1209 required to UTF-8-encode the Unicode character code point. C<UTF8SKIP()>
1210 is useful for example for iterating over the characters of a UTF-8
1211 encoded buffer; C<UNISKIP()> is useful, for example, in computing
1212 the size required for a UTF-8 encoded buffer.
1216 C<utf8_distance(a, b)> will tell the distance in characters between the
1217 two pointers pointing to the same UTF-8 encoded buffer.
1221 C<utf8_hop(s, off)> will return a pointer to an UTF-8 encoded buffer
1222 that is C<off> (positive or negative) Unicode characters displaced
1223 from the UTF-8 buffer C<s>. Be careful not to overstep the buffer:
1224 C<utf8_hop()> will merrily run off the end or the beginning of the
1225 buffer if told to do so.
1229 C<pv_uni_display(dsv, spv, len, pvlim, flags)> and
1230 C<sv_uni_display(dsv, ssv, pvlim, flags)> are useful for debugging the
1231 output of Unicode strings and scalars. By default they are useful
1232 only for debugging--they display B<all> characters as hexadecimal code
1233 points--but with the flags C<UNI_DISPLAY_ISPRINT>,
1234 C<UNI_DISPLAY_BACKSLASH>, and C<UNI_DISPLAY_QQ> you can make the
1235 output more readable.
1239 C<ibcmp_utf8(s1, pe1, u1, l1, u1, s2, pe2, l2, u2)> can be used to
1240 compare two strings case-insensitively in Unicode. For case-sensitive
1241 comparisons you can just use C<memEQ()> and C<memNE()> as usual.
1245 For more information, see L<perlapi>, and F<utf8.c> and F<utf8.h>
1246 in the Perl source code distribution.
1250 =head2 Interaction with Locales
1252 Use of locales with Unicode data may lead to odd results. Currently,
1253 Perl attempts to attach 8-bit locale info to characters in the range
1254 0..255, but this technique is demonstrably incorrect for locales that
1255 use characters above that range when mapped into Unicode. Perl's
1256 Unicode support will also tend to run slower. Use of locales with
1257 Unicode is discouraged.
1259 =head2 Interaction with Extensions
1261 When Perl exchanges data with an extension, the extension should be
1262 able to understand the UTF-8 flag and act accordingly. If the
1263 extension doesn't know about the flag, it's likely that the extension
1264 will return incorrectly-flagged data.
1266 So if you're working with Unicode data, consult the documentation of
1267 every module you're using if there are any issues with Unicode data
1268 exchange. If the documentation does not talk about Unicode at all,
1269 suspect the worst and probably look at the source to learn how the
1270 module is implemented. Modules written completely in Perl shouldn't
1271 cause problems. Modules that directly or indirectly access code written
1272 in other programming languages are at risk.
1274 For affected functions, the simple strategy to avoid data corruption is
1275 to always make the encoding of the exchanged data explicit. Choose an
1276 encoding that you know the extension can handle. Convert arguments passed
1277 to the extensions to that encoding and convert results back from that
1278 encoding. Write wrapper functions that do the conversions for you, so
1279 you can later change the functions when the extension catches up.
1281 To provide an example, let's say the popular Foo::Bar::escape_html
1282 function doesn't deal with Unicode data yet. The wrapper function
1283 would convert the argument to raw UTF-8 and convert the result back to
1284 Perl's internal representation like so:
1286 sub my_escape_html ($) {
1288 return unless defined $what;
1289 Encode::decode_utf8(Foo::Bar::escape_html(Encode::encode_utf8($what)));
1292 Sometimes, when the extension does not convert data but just stores
1293 and retrieves them, you will be in a position to use the otherwise
1294 dangerous Encode::_utf8_on() function. Let's say the popular
1295 C<Foo::Bar> extension, written in C, provides a C<param> method that
1296 lets you store and retrieve data according to these prototypes:
1298 $self->param($name, $value); # set a scalar
1299 $value = $self->param($name); # retrieve a scalar
1301 If it does not yet provide support for any encoding, one could write a
1302 derived class with such a C<param> method:
1305 my($self,$name,$value) = @_;
1306 utf8::upgrade($name); # make sure it is UTF-8 encoded
1308 utf8::upgrade($value); # make sure it is UTF-8 encoded
1309 return $self->SUPER::param($name,$value);
1311 my $ret = $self->SUPER::param($name);
1312 Encode::_utf8_on($ret); # we know, it is UTF-8 encoded
1317 Some extensions provide filters on data entry/exit points, such as
1318 DB_File::filter_store_key and family. Look out for such filters in
1319 the documentation of your extensions, they can make the transition to
1320 Unicode data much easier.
1324 Some functions are slower when working on UTF-8 encoded strings than
1325 on byte encoded strings. All functions that need to hop over
1326 characters such as length(), substr() or index(), or matching regular
1327 expressions can work B<much> faster when the underlying data are
1330 In Perl 5.8.0 the slowness was often quite spectacular; in Perl 5.8.1
1331 a caching scheme was introduced which will hopefully make the slowness
1332 somewhat less spectacular, at least for some operations. In general,
1333 operations with UTF-8 encoded strings are still slower. As an example,
1334 the Unicode properties (character classes) like C<\p{Nd}> are known to
1335 be quite a bit slower (5-20 times) than their simpler counterparts
1336 like C<\d> (then again, there 268 Unicode characters matching C<Nd>
1337 compared with the 10 ASCII characters matching C<d>).
1339 =head2 Porting code from perl-5.6.X
1341 Perl 5.8 has a different Unicode model from 5.6. In 5.6 the programmer
1342 was required to use the C<utf8> pragma to declare that a given scope
1343 expected to deal with Unicode data and had to make sure that only
1344 Unicode data were reaching that scope. If you have code that is
1345 working with 5.6, you will need some of the following adjustments to
1346 your code. The examples are written such that the code will continue
1347 to work under 5.6, so you should be safe to try them out.
1353 A filehandle that should read or write UTF-8
1356 binmode $fh, ":utf8";
1361 A scalar that is going to be passed to some extension
1363 Be it Compress::Zlib, Apache::Request or any extension that has no
1364 mention of Unicode in the manpage, you need to make sure that the
1365 UTF-8 flag is stripped off. Note that at the time of this writing
1366 (October 2002) the mentioned modules are not UTF-8-aware. Please
1367 check the documentation to verify if this is still true.
1371 $val = Encode::encode_utf8($val); # make octets
1376 A scalar we got back from an extension
1378 If you believe the scalar comes back as UTF-8, you will most likely
1379 want the UTF-8 flag restored:
1383 $val = Encode::decode_utf8($val);
1388 Same thing, if you are really sure it is UTF-8
1392 Encode::_utf8_on($val);
1397 A wrapper for fetchrow_array and fetchrow_hashref
1399 When the database contains only UTF-8, a wrapper function or method is
1400 a convenient way to replace all your fetchrow_array and
1401 fetchrow_hashref calls. A wrapper function will also make it easier to
1402 adapt to future enhancements in your database driver. Note that at the
1403 time of this writing (October 2002), the DBI has no standardized way
1404 to deal with UTF-8 data. Please check the documentation to verify if
1408 my($self, $sth, $what) = @_; # $what is one of fetchrow_{array,hashref}
1414 my @arr = $sth->$what;
1416 defined && /[^\000-\177]/ && Encode::_utf8_on($_);
1420 my $ret = $sth->$what;
1422 for my $k (keys %$ret) {
1423 defined && /[^\000-\177]/ && Encode::_utf8_on($_) for $ret->{$k};
1427 defined && /[^\000-\177]/ && Encode::_utf8_on($_) for $ret;
1437 A large scalar that you know can only contain ASCII
1439 Scalars that contain only ASCII and are marked as UTF-8 are sometimes
1440 a drag to your program. If you recognize such a situation, just remove
1443 utf8::downgrade($val) if $] > 5.007;
1449 L<perluniintro>, L<encoding>, L<Encode>, L<open>, L<utf8>, L<bytes>,
1450 L<perlretut>, L<perlvar/"${^UNICODE}">