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>.
47 =head2 Byte and Character Semantics
49 Beginning with version 5.6, Perl uses logically-wide characters to
50 represent strings internally.
52 In future, Perl-level operations will be expected to work with
53 characters rather than bytes.
55 However, as an interim compatibility measure, Perl aims to
56 provide a safe migration path from byte semantics to character
57 semantics for programs. For operations where Perl can unambiguously
58 decide that the input data are characters, Perl switches to
59 character semantics. For operations where this determination cannot
60 be made without additional information from the user, Perl decides in
61 favor of compatibility and chooses to use byte semantics.
63 This behavior preserves compatibility with earlier versions of Perl,
64 which allowed byte semantics in Perl operations only if
65 none of the program's inputs were marked as being as source of Unicode
66 character data. Such data may come from filehandles, from calls to
67 external programs, from information provided by the system (such as %ENV),
68 or from literals and constants in the source text.
70 The C<bytes> pragma will always, regardless of platform, force byte
71 semantics in a particular lexical scope. See L<bytes>.
73 The C<utf8> pragma is primarily a compatibility device that enables
74 recognition of UTF-(8|EBCDIC) in literals encountered by the parser.
75 Note that this pragma is only required while Perl defaults to byte
76 semantics; when character semantics become the default, this pragma
77 may become a no-op. See L<utf8>.
79 Unless explicitly stated, Perl operators use character semantics
80 for Unicode data and byte semantics for non-Unicode data.
81 The decision to use character semantics is made transparently. If
82 input data comes from a Unicode source--for example, if a character
83 encoding layer is added to a filehandle or a literal Unicode
84 string constant appears in a program--character semantics apply.
85 Otherwise, byte semantics are in effect. The C<bytes> pragma should
86 be used to force byte semantics on Unicode data.
88 If strings operating under byte semantics and strings with Unicode
89 character data are concatenated, the new string will be upgraded to
90 I<ISO 8859-1 (Latin-1)>, even if the old Unicode string used EBCDIC.
91 This translation is done without regard to the system's native 8-bit
92 encoding, so to change this for systems with non-Latin-1 and
93 non-EBCDIC native encodings use the C<encoding> pragma. See
96 Under character semantics, many operations that formerly operated on
97 bytes now operate on characters. A character in Perl is
98 logically just a number ranging from 0 to 2**31 or so. Larger
99 characters may encode into longer sequences of bytes internally, but
100 this internal detail is mostly hidden for Perl code.
101 See L<perluniintro> for more.
103 =head2 Effects of Character Semantics
105 Character semantics have the following effects:
111 Strings--including hash keys--and regular expression patterns may
112 contain characters that have an ordinal value larger than 255.
114 If you use a Unicode editor to edit your program, Unicode characters
115 may occur directly within the literal strings in one of the various
116 Unicode encodings (UTF-8, UTF-EBCDIC, UCS-2, etc.), but will be recognized
117 as such and converted to Perl's internal representation only if the
118 appropriate L<encoding> is specified.
120 Unicode characters can also be added to a string by using the
121 C<\x{...}> notation. The Unicode code for the desired character, in
122 hexadecimal, should be placed in the braces. For instance, a smiley
123 face is C<\x{263A}>. This encoding scheme only works for characters
124 with a code of 0x100 or above.
128 use charnames ':full';
130 you can use the C<\N{...}> notation and put the official Unicode
131 character name within the braces, such as C<\N{WHITE SMILING FACE}>.
136 If an appropriate L<encoding> is specified, identifiers within the
137 Perl script may contain Unicode alphanumeric characters, including
138 ideographs. Perl does not currently attempt to canonicalize variable
143 Regular expressions match characters instead of bytes. "." matches
144 a character instead of a byte. The C<\C> pattern is provided to force
145 a match a single byte--a C<char> in C, hence C<\C>.
149 Character classes in regular expressions match characters instead of
150 bytes and match against the character properties specified in the
151 Unicode properties database. C<\w> can be used to match a Japanese
152 ideograph, for instance.
156 Named Unicode properties, scripts, and block ranges may be used like
157 character classes via the C<\p{}> "matches property" construct and
158 the C<\P{}> negation, "doesn't match property".
160 For instance, C<\p{Lu}> matches any character with the Unicode "Lu"
161 (Letter, uppercase) property, while C<\p{M}> matches any character
162 with an "M" (mark--accents and such) property. Brackets are not
163 required for single letter properties, so C<\p{M}> is equivalent to
164 C<\pM>. Many predefined properties are available, such as
165 C<\p{Mirrored}> and C<\p{Tibetan}>.
167 The official Unicode script and block names have spaces and dashes as
168 separators, but for convenience you can use dashes, spaces, or
169 underbars, and case is unimportant. It is recommended, however, that
170 for consistency you use the following naming: the official Unicode
171 script, property, or block name (see below for the additional rules
172 that apply to block names) with whitespace and dashes removed, and the
173 words "uppercase-first-lowercase-rest". C<Latin-1 Supplement> thus
174 becomes C<Latin1Supplement>.
176 You can also use negation in both C<\p{}> and C<\P{}> by introducing a caret
177 (^) between the first brace and the property name: C<\p{^Tamil}> is
178 equal to C<\P{Tamil}>.
180 Here are the basic Unicode General Category properties, followed by their
181 long form. You can use either; C<\p{Lu}> and C<\p{UppercaseLetter}>,
182 for instance, are identical.
204 Pc ConnectorPunctuation
208 Pi InitialPunctuation
209 (may behave like Ps or Pe depending on usage)
211 (may behave like Ps or Pe depending on usage)
223 Zp ParagraphSeparator
228 Cs Surrogate (not usable)
232 Single-letter properties match all characters in any of the
233 two-letter sub-properties starting with the same letter.
234 C<L&> is a special case, which is an alias for C<Ll>, C<Lu>, and C<Lt>.
236 Because Perl hides the need for the user to understand the internal
237 representation of Unicode characters, there is no need to implement
238 the somewhat messy concept of surrogates. C<Cs> is therefore not
241 Because scripts differ in their directionality--Hebrew is
242 written right to left, for example--Unicode supplies these properties:
247 BidiLRE Left-to-Right Embedding
248 BidiLRO Left-to-Right Override
250 BidiAL Right-to-Left Arabic
251 BidiRLE Right-to-Left Embedding
252 BidiRLO Right-to-Left Override
253 BidiPDF Pop Directional Format
254 BidiEN European Number
255 BidiES European Number Separator
256 BidiET European Number Terminator
258 BidiCS Common Number Separator
259 BidiNSM Non-Spacing Mark
260 BidiBN Boundary Neutral
261 BidiB Paragraph Separator
262 BidiS Segment Separator
264 BidiON Other Neutrals
266 For example, C<\p{BidiR}> matches characters that are normally
267 written right to left.
273 The script names which can be used by C<\p{...}> and C<\P{...}>,
274 such as in C<\p{Latin}> or C<\p{Cyrillic}>, are as follows:
321 Extended property classes can supplement the basic
322 properties, defined by the F<PropList> Unicode database:
337 LogicalOrderException
338 NoncharacterCodePoint
340 OtherDefaultIgnorableCodePoint
352 and there are further derived properties:
354 Alphabetic Lu + Ll + Lt + Lm + Lo + OtherAlphabetic
355 Lowercase Ll + OtherLowercase
356 Uppercase Lu + OtherUppercase
359 ID_Start Lu + Ll + Lt + Lm + Lo + Nl
360 ID_Continue ID_Start + Mn + Mc + Nd + Pc
363 Assigned Any non-Cn character (i.e. synonym for \P{Cn})
364 Unassigned Synonym for \p{Cn}
365 Common Any character (or unassigned code point)
366 not explicitly assigned to a script
368 For backward compatibility (with Perl 5.6), all properties mentioned
369 so far may have C<Is> prepended to their name, so C<\P{IsLu}>, for
370 example, is equal to C<\P{Lu}>.
374 In addition to B<scripts>, Unicode also defines B<blocks> of
375 characters. The difference between scripts and blocks is that the
376 concept of scripts is closer to natural languages, while the concept
377 of blocks is more of an artificial grouping based on groups of 256
378 Unicode characters. For example, the C<Latin> script contains letters
379 from many blocks but does not contain all the characters from those
380 blocks. It does not, for example, contain digits, because digits are
381 shared across many scripts. Digits and similar groups, like
382 punctuation, are in a category called C<Common>.
384 For more about scripts, see the UTR #24:
386 http://www.unicode.org/unicode/reports/tr24/
388 For more about blocks, see:
390 http://www.unicode.org/Public/UNIDATA/Blocks.txt
392 Block names are given with the C<In> prefix. For example, the
393 Katakana block is referenced via C<\p{InKatakana}>. The C<In>
394 prefix may be omitted if there is no naming conflict with a script
395 or any other property, but it is recommended that C<In> always be used
396 for block tests to avoid confusion.
398 These block names are supported:
400 InAlphabeticPresentationForms
402 InArabicPresentationFormsA
403 InArabicPresentationFormsB
414 InByzantineMusicalSymbols
416 InCJKCompatibilityForms
417 InCJKCompatibilityIdeographs
418 InCJKCompatibilityIdeographsSupplement
419 InCJKRadicalsSupplement
420 InCJKSymbolsAndPunctuation
421 InCJKUnifiedIdeographs
422 InCJKUnifiedIdeographsExtensionA
423 InCJKUnifiedIdeographsExtensionB
425 InCombiningDiacriticalMarks
426 InCombiningDiacriticalMarksforSymbols
431 InCyrillicSupplementary
435 InEnclosedAlphanumerics
436 InEnclosedCJKLettersAndMonths
446 InHalfwidthAndFullwidthForms
447 InHangulCompatibilityJamo
452 InHighPrivateUseSurrogates
456 InIdeographicDescriptionCharacters
461 InKatakanaPhoneticExtensions
466 InLatinExtendedAdditional
471 InMathematicalAlphanumericSymbols
472 InMathematicalOperators
473 InMiscellaneousMathematicalSymbolsA
474 InMiscellaneousMathematicalSymbolsB
475 InMiscellaneousSymbols
476 InMiscellaneousTechnical
483 InOpticalCharacterRecognition
489 InSpacingModifierLetters
491 InSuperscriptsAndSubscripts
492 InSupplementalArrowsA
493 InSupplementalArrowsB
494 InSupplementalMathematicalOperators
495 InSupplementaryPrivateUseAreaA
496 InSupplementaryPrivateUseAreaB
506 InUnifiedCanadianAboriginalSyllabics
515 The special pattern C<\X> matches any extended Unicode
516 sequence--"a combining character sequence" in Standardese--where the
517 first character is a base character and subsequent characters are mark
518 characters that apply to the base character. C<\X> is equivalent to
523 The C<tr///> operator translates characters instead of bytes. Note
524 that the C<tr///CU> functionality has been removed. For similar
525 functionality see pack('U0', ...) and pack('C0', ...).
529 Case translation operators use the Unicode case translation tables
530 when character input is provided. Note that C<uc()>, or C<\U> in
531 interpolated strings, translates to uppercase, while C<ucfirst>,
532 or C<\u> in interpolated strings, translates to titlecase in languages
533 that make the distinction.
537 Most operators that deal with positions or lengths in a string will
538 automatically switch to using character positions, including
539 C<chop()>, C<substr()>, C<pos()>, C<index()>, C<rindex()>,
540 C<sprintf()>, C<write()>, and C<length()>. Operators that
541 specifically do not switch include C<vec()>, C<pack()>, and
542 C<unpack()>. Operators that really don't care include C<chomp()>,
543 operators that treats strings as a bucket of bits such as C<sort()>,
544 and operators dealing with filenames.
548 The C<pack()>/C<unpack()> letters C<c> and C<C> do I<not> change,
549 since they are often used for byte-oriented formats. Again, think
550 C<char> in the C language.
552 There is a new C<U> specifier that converts between Unicode characters
557 The C<chr()> and C<ord()> functions work on characters, similar to
558 C<pack("U")> and C<unpack("U")>, I<not> C<pack("C")> and
559 C<unpack("C")>. C<pack("C")> and C<unpack("C")> are methods for
560 emulating byte-oriented C<chr()> and C<ord()> on Unicode strings.
561 While these methods reveal the internal encoding of Unicode strings,
562 that is not something one normally needs to care about at all.
566 The bit string operators, C<& | ^ ~>, can operate on character data.
567 However, for backward compatibility, such as when using bit string
568 operations when characters are all less than 256 in ordinal value, one
569 should not use C<~> (the bit complement) with characters of both
570 values less than 256 and values greater than 256. Most importantly,
571 DeMorgan's laws (C<~($x|$y) eq ~$x&~$y> and C<~($x&$y) eq ~$x|~$y>)
572 will not hold. The reason for this mathematical I<faux pas> is that
573 the complement cannot return B<both> the 8-bit (byte-wide) bit
574 complement B<and> the full character-wide bit complement.
578 lc(), uc(), lcfirst(), and ucfirst() work for the following cases:
584 the case mapping is from a single Unicode character to another
585 single Unicode character, or
589 the case mapping is from a single Unicode character to more
590 than one Unicode character.
594 Things to do with locales (Lithuanian, Turkish, Azeri) do B<not> work
595 since Perl does not understand the concept of Unicode locales.
597 See the Unicode Technical Report #21, Case Mappings, for more details.
605 And finally, C<scalar reverse()> reverses by character rather than by byte.
609 =head2 User-Defined Character Properties
611 You can define your own character properties by defining subroutines
612 whose names begin with "In" or "Is". The subroutines must be defined
613 in the C<main> package. The user-defined properties can be used in the
614 regular expression C<\p> and C<\P> constructs. Note that the effect
615 is compile-time and immutable once defined.
617 The subroutines must return a specially-formatted string, with one
618 or more newline-separated lines. Each line must be one of the following:
624 Two hexadecimal numbers separated by horizontal whitespace (space or
625 tabular characters) denoting a range of Unicode code points to include.
629 Something to include, prefixed by "+": a built-in character
630 property (prefixed by "utf8::"), to represent all the characters in that
631 property; two hexadecimal code points for a range; or a single
632 hexadecimal code point.
636 Something to exclude, prefixed by "-": an existing character
637 property (prefixed by "utf8::"), for all the characters in that
638 property; two hexadecimal code points for a range; or a single
639 hexadecimal code point.
643 Something to negate, prefixed "!": an existing character
644 property (prefixed by "utf8::") for all the characters except the
645 characters in the property; two hexadecimal code points for a range;
646 or a single hexadecimal code point.
650 For example, to define a property that covers both the Japanese
651 syllabaries (hiragana and katakana), you can define
660 Imagine that the here-doc end marker is at the beginning of the line.
661 Now you can use C<\p{InKana}> and C<\P{InKana}>.
663 You could also have used the existing block property names:
672 Suppose you wanted to match only the allocated characters,
673 not the raw block ranges: in other words, you want to remove
684 The negation is useful for defining (surprise!) negated classes.
694 You can also define your own mappings to be used in the lc(),
695 lcfirst(), uc(), and ucfirst() (or their string-inlined versions).
696 The principle is the same: define subroutines in the C<main> package
697 with names like C<ToLower> (for lc() and lcfirst()), C<ToTitle> (for
698 the first character in ucfirst()), and C<ToUpper> (for uc(), and the
699 rest of the characters in ucfirst()).
701 The string returned by the subroutines needs now to be three
702 hexadecimal numbers separated by tabulators: start of the source
703 range, end of the source range, and start of the destination range.
712 defines an uc() mapping that causes only the characters "a", "b", and
713 "c" to be mapped to "A", "B", "C", all other characters will remain
716 If there is no source range to speak of, that is, the mapping is from
717 a single character to another single character, leave the end of the
718 source range empty, but the two tabulator characters are still needed.
727 defines a lc() mapping that causes only "A" to be mapped to "a", all
728 other characters will remain unchanged.
730 (For serious hackers only) If you want to introspect the default
731 mappings, you can find the data in the directory
732 C<$Config{privlib}>/F<unicore/To/>. The mapping data is returned as
733 the here-document, and the C<utf8::ToSpecFoo> are special exception
734 mappings derived from <$Config{privlib}>/F<unicore/SpecialCasing.txt>.
735 The C<Digit> and C<Fold> mappings that one can see in the directory
736 are not directly user-accessible, one can use either the
737 C<Unicode::UCD> module, or just match case-insensitively (that's when
738 the C<Fold> mapping is used).
740 A final note on the user-defined property tests and mappings: they
741 will be used only if the scalar has been marked as having Unicode
742 characters. Old byte-style strings will not be affected.
744 =head2 Character Encodings for Input and Output
748 =head2 Unicode Regular Expression Support Level
750 The following list of Unicode support for regular expressions describes
751 all the features currently supported. The references to "Level N"
752 and the section numbers refer to the Unicode Technical Report 18,
753 "Unicode Regular Expression Guidelines".
759 Level 1 - Basic Unicode Support
761 2.1 Hex Notation - done [1]
762 Named Notation - done [2]
763 2.2 Categories - done [3][4]
764 2.3 Subtraction - MISSING [5][6]
765 2.4 Simple Word Boundaries - done [7]
766 2.5 Simple Loose Matches - done [8]
767 2.6 End of Line - MISSING [9][10]
771 [ 3] . \p{...} \P{...}
772 [ 4] now scripts (see UTR#24 Script Names) in addition to blocks
774 [ 6] can use regular expression look-ahead [a]
775 or user-defined character properties [b] to emulate subtraction
776 [ 7] include Letters in word characters
777 [ 8] note that Perl does Full case-folding in matching, not Simple:
778 for example U+1F88 is equivalent with U+1F00 U+03B9,
779 not with 1F80. This difference matters for certain Greek
780 capital letters with certain modifiers: the Full case-folding
781 decomposes the letter, while the Simple case-folding would map
782 it to a single character.
783 [ 9] see UTR#13 Unicode Newline Guidelines
784 [10] should do ^ and $ also on \x{85}, \x{2028} and \x{2029}
785 (should also affect <>, $., and script line numbers)
786 (the \x{85}, \x{2028} and \x{2029} do match \s)
788 [a] You can mimic class subtraction using lookahead.
789 For example, what TR18 might write as
791 [{Greek}-[{UNASSIGNED}]]
793 in Perl can be written as:
795 (?!\p{Unassigned})\p{InGreekAndCoptic}
796 (?=\p{Assigned})\p{InGreekAndCoptic}
798 But in this particular example, you probably really want
802 which will match assigned characters known to be part of the Greek script.
804 [b] See L</"User-Defined Character Properties">.
808 Level 2 - Extended Unicode Support
810 3.1 Surrogates - MISSING [11]
811 3.2 Canonical Equivalents - MISSING [12][13]
812 3.3 Locale-Independent Graphemes - MISSING [14]
813 3.4 Locale-Independent Words - MISSING [15]
814 3.5 Locale-Independent Loose Matches - MISSING [16]
816 [11] Surrogates are solely a UTF-16 concept and Perl's internal
817 representation is UTF-8. The Encode module does UTF-16, though.
818 [12] see UTR#15 Unicode Normalization
819 [13] have Unicode::Normalize but not integrated to regexes
820 [14] have \X but at this level . should equal that
821 [15] need three classes, not just \w and \W
822 [16] see UTR#21 Case Mappings
826 Level 3 - Locale-Sensitive Support
828 4.1 Locale-Dependent Categories - MISSING
829 4.2 Locale-Dependent Graphemes - MISSING [16][17]
830 4.3 Locale-Dependent Words - MISSING
831 4.4 Locale-Dependent Loose Matches - MISSING
832 4.5 Locale-Dependent Ranges - MISSING
834 [16] see UTR#10 Unicode Collation Algorithms
835 [17] have Unicode::Collate but not integrated to regexes
839 =head2 Unicode Encodings
841 Unicode characters are assigned to I<code points>, which are abstract
842 numbers. To use these numbers, various encodings are needed.
850 UTF-8 is a variable-length (1 to 6 bytes, current character allocations
851 require 4 bytes), byte-order independent encoding. For ASCII (and we
852 really do mean 7-bit ASCII, not another 8-bit encoding), UTF-8 is
855 The following table is from Unicode 3.2.
857 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
859 U+0000..U+007F 00..7F
860 U+0080..U+07FF C2..DF 80..BF
861 U+0800..U+0FFF E0 A0..BF 80..BF
862 U+1000..U+CFFF E1..EC 80..BF 80..BF
863 U+D000..U+D7FF ED 80..9F 80..BF
864 U+D800..U+DFFF ******* ill-formed *******
865 U+E000..U+FFFF EE..EF 80..BF 80..BF
866 U+10000..U+3FFFF F0 90..BF 80..BF 80..BF
867 U+40000..U+FFFFF F1..F3 80..BF 80..BF 80..BF
868 U+100000..U+10FFFF F4 80..8F 80..BF 80..BF
870 Note the C<A0..BF> in C<U+0800..U+0FFF>, the C<80..9F> in
871 C<U+D000...U+D7FF>, the C<90..B>F in C<U+10000..U+3FFFF>, and the
872 C<80...8F> in C<U+100000..U+10FFFF>. The "gaps" are caused by legal
873 UTF-8 avoiding non-shortest encodings: it is technically possible to
874 UTF-8-encode a single code point in different ways, but that is
875 explicitly forbidden, and the shortest possible encoding should always
876 be used. So that's what Perl does.
878 Another way to look at it is via bits:
880 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
883 00000bbbbbaaaaaa 110bbbbb 10aaaaaa
884 ccccbbbbbbaaaaaa 1110cccc 10bbbbbb 10aaaaaa
885 00000dddccccccbbbbbbaaaaaa 11110ddd 10cccccc 10bbbbbb 10aaaaaa
887 As you can see, the continuation bytes all begin with C<10>, and the
888 leading bits of the start byte tell how many bytes the are in the
895 Like UTF-8 but EBCDIC-safe, in the way that UTF-8 is ASCII-safe.
899 UTF-16, UTF-16BE, UTF16-LE, Surrogates, and BOMs (Byte Order Marks)
901 The followings items are mostly for reference and general Unicode
902 knowledge, Perl doesn't use these constructs internally.
904 UTF-16 is a 2 or 4 byte encoding. The Unicode code points
905 C<U+0000..U+FFFF> are stored in a single 16-bit unit, and the code
906 points C<U+10000..U+10FFFF> in two 16-bit units. The latter case is
907 using I<surrogates>, the first 16-bit unit being the I<high
908 surrogate>, and the second being the I<low surrogate>.
910 Surrogates are code points set aside to encode the C<U+10000..U+10FFFF>
911 range of Unicode code points in pairs of 16-bit units. The I<high
912 surrogates> are the range C<U+D800..U+DBFF>, and the I<low surrogates>
913 are the range C<U+DC00..U+DFFF>. The surrogate encoding is
915 $hi = ($uni - 0x10000) / 0x400 + 0xD800;
916 $lo = ($uni - 0x10000) % 0x400 + 0xDC00;
920 $uni = 0x10000 + ($hi - 0xD800) * 0x400 + ($lo - 0xDC00);
922 If you try to generate surrogates (for example by using chr()), you
923 will get a warning if warnings are turned on, because those code
924 points are not valid for a Unicode character.
926 Because of the 16-bitness, UTF-16 is byte-order dependent. UTF-16
927 itself can be used for in-memory computations, but if storage or
928 transfer is required either UTF-16BE (big-endian) or UTF-16LE
929 (little-endian) encodings must be chosen.
931 This introduces another problem: what if you just know that your data
932 is UTF-16, but you don't know which endianness? Byte Order Marks, or
933 BOMs, are a solution to this. A special character has been reserved
934 in Unicode to function as a byte order marker: the character with the
935 code point C<U+FEFF> is the BOM.
937 The trick is that if you read a BOM, you will know the byte order,
938 since if it was written on a big-endian platform, you will read the
939 bytes C<0xFE 0xFF>, but if it was written on a little-endian platform,
940 you will read the bytes C<0xFF 0xFE>. (And if the originating platform
941 was writing in UTF-8, you will read the bytes C<0xEF 0xBB 0xBF>.)
943 The way this trick works is that the character with the code point
944 C<U+FFFE> is guaranteed not to be a valid Unicode character, so the
945 sequence of bytes C<0xFF 0xFE> is unambiguously "BOM, represented in
946 little-endian format" and cannot be C<U+FFFE>, represented in big-endian
951 UTF-32, UTF-32BE, UTF32-LE
953 The UTF-32 family is pretty much like the UTF-16 family, expect that
954 the units are 32-bit, and therefore the surrogate scheme is not
955 needed. The BOM signatures will be C<0x00 0x00 0xFE 0xFF> for BE and
956 C<0xFF 0xFE 0x00 0x00> for LE.
962 Encodings defined by the ISO 10646 standard. UCS-2 is a 16-bit
963 encoding. Unlike UTF-16, UCS-2 is not extensible beyond C<U+FFFF>,
964 because it does not use surrogates. UCS-4 is a 32-bit encoding,
965 functionally identical to UTF-32.
971 A seven-bit safe (non-eight-bit) encoding, which is useful if the
972 transport or storage is not eight-bit safe. Defined by RFC 2152.
976 =head2 Security Implications of Unicode
984 Unfortunately, the specification of UTF-8 leaves some room for
985 interpretation of how many bytes of encoded output one should generate
986 from one input Unicode character. Strictly speaking, the shortest
987 possible sequence of UTF-8 bytes should be generated,
988 because otherwise there is potential for an input buffer overflow at
989 the receiving end of a UTF-8 connection. Perl always generates the
990 shortest length UTF-8, and with warnings on Perl will warn about
991 non-shortest length UTF-8 along with other malformations, such as the
992 surrogates, which are not real Unicode code points.
996 Regular expressions behave slightly differently between byte data and
997 character (Unicode) data. For example, the "word character" character
998 class C<\w> will work differently depending on if data is eight-bit bytes
1001 In the first case, the set of C<\w> characters is either small--the
1002 default set of alphabetic characters, digits, and the "_"--or, if you
1003 are using a locale (see L<perllocale>), the C<\w> might contain a few
1004 more letters according to your language and country.
1006 In the second case, the C<\w> set of characters is much, much larger.
1007 Most importantly, even in the set of the first 256 characters, it will
1008 probably match different characters: unlike most locales, which are
1009 specific to a language and country pair, Unicode classifies all the
1010 characters that are letters I<somewhere> as C<\w>. For example, your
1011 locale might not think that LATIN SMALL LETTER ETH is a letter (unless
1012 you happen to speak Icelandic), but Unicode does.
1014 As discussed elsewhere, Perl has one foot (two hooves?) planted in
1015 each of two worlds: the old world of bytes and the new world of
1016 characters, upgrading from bytes to characters when necessary.
1017 If your legacy code does not explicitly use Unicode, no automatic
1018 switch-over to characters should happen. Characters shouldn't get
1019 downgraded to bytes, either. It is possible to accidentally mix bytes
1020 and characters, however (see L<perluniintro>), in which case C<\w> in
1021 regular expressions might start behaving differently. Review your
1022 code. Use warnings and the C<strict> pragma.
1026 =head2 Unicode in Perl on EBCDIC
1028 The way Unicode is handled on EBCDIC platforms is still
1029 experimental. On such platforms, references to UTF-8 encoding in this
1030 document and elsewhere should be read as meaning the UTF-EBCDIC
1031 specified in Unicode Technical Report 16, unless ASCII vs. EBCDIC issues
1032 are specifically discussed. There is no C<utfebcdic> pragma or
1033 ":utfebcdic" layer; rather, "utf8" and ":utf8" are reused to mean
1034 the platform's "natural" 8-bit encoding of Unicode. See L<perlebcdic>
1035 for more discussion of the issues.
1039 Usually locale settings and Unicode do not affect each other, but
1040 there are a couple of exceptions:
1046 If your locale environment variables (LC_ALL, LC_CTYPE, LANG) contain
1047 the strings 'UTF-8' or 'UTF8' (matched case-insensitively) B<and> you
1048 enable using UTF-8 either by using the C<-C> command line switch or
1049 setting the PERL_UNICODE environment variable to be, C<"L">, an empty
1050 string, (see L<perlrun> and the documentation for the C<-C> switch for
1051 more information about the possible values), then the default
1052 encodings of your STDIN, STDOUT, and STDERR, and of B<any subsequent
1053 file open>, are considered to be UTF-8. See L<perluniintro>,
1054 L<perlfunc/open>, and L<open> for more information. The magic
1055 variable C<${^UNICODE}> will also be set.
1059 Perl tries really hard to work both with Unicode and the old
1060 byte-oriented world. Most often this is nice, but sometimes Perl's
1061 straddling of the proverbial fence causes problems.
1065 =head2 Using Unicode in XS
1067 If you want to handle Perl Unicode in XS extensions, you may find the
1068 following C APIs useful. See also L<perlguts/"Unicode Support"> for an
1069 explanation about Unicode at the XS level, and L<perlapi> for the API
1076 C<DO_UTF8(sv)> returns true if the C<UTF8> flag is on and the bytes
1077 pragma is not in effect. C<SvUTF8(sv)> returns true is the C<UTF8>
1078 flag is on; the bytes pragma is ignored. The C<UTF8> flag being on
1079 does B<not> mean that there are any characters of code points greater
1080 than 255 (or 127) in the scalar or that there are even any characters
1081 in the scalar. What the C<UTF8> flag means is that the sequence of
1082 octets in the representation of the scalar is the sequence of UTF-8
1083 encoded code points of the characters of a string. The C<UTF8> flag
1084 being off means that each octet in this representation encodes a
1085 single character with code point 0..255 within the string. Perl's
1086 Unicode model is not to use UTF-8 until it is absolutely necessary.
1090 C<uvuni_to_utf8(buf, chr>) writes a Unicode character code point into
1091 a buffer encoding the code point as UTF-8, and returns a pointer
1092 pointing after the UTF-8 bytes.
1096 C<utf8_to_uvuni(buf, lenp)> reads UTF-8 encoded bytes from a buffer and
1097 returns the Unicode character code point and, optionally, the length of
1098 the UTF-8 byte sequence.
1102 C<utf8_length(start, end)> returns the length of the UTF-8 encoded buffer
1103 in characters. C<sv_len_utf8(sv)> returns the length of the UTF-8 encoded
1108 C<sv_utf8_upgrade(sv)> converts the string of the scalar to its UTF-8
1109 encoded form. C<sv_utf8_downgrade(sv)> does the opposite, if
1110 possible. C<sv_utf8_encode(sv)> is like sv_utf8_upgrade except that
1111 it does not set the C<UTF8> flag. C<sv_utf8_decode()> does the
1112 opposite of C<sv_utf8_encode()>. Note that none of these are to be
1113 used as general-purpose encoding or decoding interfaces: C<use Encode>
1114 for that. C<sv_utf8_upgrade()> is affected by the encoding pragma
1115 but C<sv_utf8_downgrade()> is not (since the encoding pragma is
1116 designed to be a one-way street).
1120 C<is_utf8_char(s)> returns true if the pointer points to a valid UTF-8
1125 C<is_utf8_string(buf, len)> returns true if C<len> bytes of the buffer
1130 C<UTF8SKIP(buf)> will return the number of bytes in the UTF-8 encoded
1131 character in the buffer. C<UNISKIP(chr)> will return the number of bytes
1132 required to UTF-8-encode the Unicode character code point. C<UTF8SKIP()>
1133 is useful for example for iterating over the characters of a UTF-8
1134 encoded buffer; C<UNISKIP()> is useful, for example, in computing
1135 the size required for a UTF-8 encoded buffer.
1139 C<utf8_distance(a, b)> will tell the distance in characters between the
1140 two pointers pointing to the same UTF-8 encoded buffer.
1144 C<utf8_hop(s, off)> will return a pointer to an UTF-8 encoded buffer
1145 that is C<off> (positive or negative) Unicode characters displaced
1146 from the UTF-8 buffer C<s>. Be careful not to overstep the buffer:
1147 C<utf8_hop()> will merrily run off the end or the beginning of the
1148 buffer if told to do so.
1152 C<pv_uni_display(dsv, spv, len, pvlim, flags)> and
1153 C<sv_uni_display(dsv, ssv, pvlim, flags)> are useful for debugging the
1154 output of Unicode strings and scalars. By default they are useful
1155 only for debugging--they display B<all> characters as hexadecimal code
1156 points--but with the flags C<UNI_DISPLAY_ISPRINT>,
1157 C<UNI_DISPLAY_BACKSLASH>, and C<UNI_DISPLAY_QQ> you can make the
1158 output more readable.
1162 C<ibcmp_utf8(s1, pe1, u1, l1, u1, s2, pe2, l2, u2)> can be used to
1163 compare two strings case-insensitively in Unicode. For case-sensitive
1164 comparisons you can just use C<memEQ()> and C<memNE()> as usual.
1168 For more information, see L<perlapi>, and F<utf8.c> and F<utf8.h>
1169 in the Perl source code distribution.
1173 =head2 Interaction with Locales
1175 Use of locales with Unicode data may lead to odd results. Currently,
1176 Perl attempts to attach 8-bit locale info to characters in the range
1177 0..255, but this technique is demonstrably incorrect for locales that
1178 use characters above that range when mapped into Unicode. Perl's
1179 Unicode support will also tend to run slower. Use of locales with
1180 Unicode is discouraged.
1182 =head2 Interaction with Extensions
1184 When Perl exchanges data with an extension, the extension should be
1185 able to understand the UTF-8 flag and act accordingly. If the
1186 extension doesn't know about the flag, it's likely that the extension
1187 will return incorrectly-flagged data.
1189 So if you're working with Unicode data, consult the documentation of
1190 every module you're using if there are any issues with Unicode data
1191 exchange. If the documentation does not talk about Unicode at all,
1192 suspect the worst and probably look at the source to learn how the
1193 module is implemented. Modules written completely in Perl shouldn't
1194 cause problems. Modules that directly or indirectly access code written
1195 in other programming languages are at risk.
1197 For affected functions, the simple strategy to avoid data corruption is
1198 to always make the encoding of the exchanged data explicit. Choose an
1199 encoding that you know the extension can handle. Convert arguments passed
1200 to the extensions to that encoding and convert results back from that
1201 encoding. Write wrapper functions that do the conversions for you, so
1202 you can later change the functions when the extension catches up.
1204 To provide an example, let's say the popular Foo::Bar::escape_html
1205 function doesn't deal with Unicode data yet. The wrapper function
1206 would convert the argument to raw UTF-8 and convert the result back to
1207 Perl's internal representation like so:
1209 sub my_escape_html ($) {
1211 return unless defined $what;
1212 Encode::decode_utf8(Foo::Bar::escape_html(Encode::encode_utf8($what)));
1215 Sometimes, when the extension does not convert data but just stores
1216 and retrieves them, you will be in a position to use the otherwise
1217 dangerous Encode::_utf8_on() function. Let's say the popular
1218 C<Foo::Bar> extension, written in C, provides a C<param> method that
1219 lets you store and retrieve data according to these prototypes:
1221 $self->param($name, $value); # set a scalar
1222 $value = $self->param($name); # retrieve a scalar
1224 If it does not yet provide support for any encoding, one could write a
1225 derived class with such a C<param> method:
1228 my($self,$name,$value) = @_;
1229 utf8::upgrade($name); # make sure it is UTF-8 encoded
1231 utf8::upgrade($value); # make sure it is UTF-8 encoded
1232 return $self->SUPER::param($name,$value);
1234 my $ret = $self->SUPER::param($name);
1235 Encode::_utf8_on($ret); # we know, it is UTF-8 encoded
1240 Some extensions provide filters on data entry/exit points, such as
1241 DB_File::filter_store_key and family. Look out for such filters in
1242 the documentation of your extensions, they can make the transition to
1243 Unicode data much easier.
1247 Some functions are slower when working on UTF-8 encoded strings than
1248 on byte encoded strings. All functions that need to hop over
1249 characters such as length(), substr() or index() can work B<much>
1250 faster when the underlying data are byte-encoded. Witness the
1251 following benchmark:
1257 our $u = our $b = "x" x $l;
1258 substr($u,0,1) = "\x{100}";
1260 LENGTH_B => q{ length($b) },
1261 LENGTH_U => q{ length($u) },
1262 SUBSTR_B => q{ substr($b, $l/4, $l/2) },
1263 SUBSTR_U => q{ substr($u, $l/4, $l/2) },
1266 Benchmark: running LENGTH_B, LENGTH_U, SUBSTR_B, SUBSTR_U for at least 2 CPU seconds...
1267 LENGTH_B: 2 wallclock secs ( 2.36 usr + 0.00 sys = 2.36 CPU) @ 5649983.05/s (n=13333960)
1268 LENGTH_U: 2 wallclock secs ( 2.11 usr + 0.00 sys = 2.11 CPU) @ 12155.45/s (n=25648)
1269 SUBSTR_B: 3 wallclock secs ( 2.16 usr + 0.00 sys = 2.16 CPU) @ 374480.09/s (n=808877)
1270 SUBSTR_U: 2 wallclock secs ( 2.11 usr + 0.00 sys = 2.11 CPU) @ 6791.00/s (n=14329)
1272 The numbers show an incredible slowness on long UTF-8 strings. You
1273 should carefully avoid using these functions in tight loops. If you
1274 want to iterate over characters, the superior coding technique would
1275 split the characters into an array instead of using substr, as the following
1282 our $u = our $b = "x" x $l;
1283 substr($u,0,1) = "\x{100}";
1285 SPLIT_B => q{ for my $c (split //, $b){} },
1286 SPLIT_U => q{ for my $c (split //, $u){} },
1287 SUBSTR_B => q{ for my $i (0..length($b)-1){my $c = substr($b,$i,1);} },
1288 SUBSTR_U => q{ for my $i (0..length($u)-1){my $c = substr($u,$i,1);} },
1291 Benchmark: running SPLIT_B, SPLIT_U, SUBSTR_B, SUBSTR_U for at least 5 CPU seconds...
1292 SPLIT_B: 6 wallclock secs ( 5.29 usr + 0.00 sys = 5.29 CPU) @ 56.14/s (n=297)
1293 SPLIT_U: 5 wallclock secs ( 5.17 usr + 0.01 sys = 5.18 CPU) @ 55.21/s (n=286)
1294 SUBSTR_B: 5 wallclock secs ( 5.34 usr + 0.00 sys = 5.34 CPU) @ 123.22/s (n=658)
1295 SUBSTR_U: 7 wallclock secs ( 6.20 usr + 0.00 sys = 6.20 CPU) @ 0.81/s (n=5)
1297 Even though the algorithm based on C<substr()> is faster than
1298 C<split()> for byte-encoded data, it pales in comparison to the speed
1299 of C<split()> when used with UTF-8 data.
1301 =head2 Porting code from perl-5.6.X
1303 Perl 5.8 has a different Unicode model from 5.6. In 5.6 the programmer
1304 was required to use the C<utf8> pragma to declare that a given scope
1305 expected to deal with Unicode data and had to make sure that only
1306 Unicode data were reaching that scope. If you have code that is
1307 working with 5.6, you will need some of the following adjustments to
1308 your code. The examples are written such that the code will continue
1309 to work under 5.6, so you should be safe to try them out.
1315 A filehandle that should read or write UTF-8
1318 binmode $fh, ":utf8";
1323 A scalar that is going to be passed to some extension
1325 Be it Compress::Zlib, Apache::Request or any extension that has no
1326 mention of Unicode in the manpage, you need to make sure that the
1327 UTF-8 flag is stripped off. Note that at the time of this writing
1328 (October 2002) the mentioned modules are not UTF-8-aware. Please
1329 check the documentation to verify if this is still true.
1333 $val = Encode::encode_utf8($val); # make octets
1338 A scalar we got back from an extension
1340 If you believe the scalar comes back as UTF-8, you will most likely
1341 want the UTF-8 flag restored:
1345 $val = Encode::decode_utf8($val);
1350 Same thing, if you are really sure it is UTF-8
1354 Encode::_utf8_on($val);
1359 A wrapper for fetchrow_array and fetchrow_hashref
1361 When the database contains only UTF-8, a wrapper function or method is
1362 a convenient way to replace all your fetchrow_array and
1363 fetchrow_hashref calls. A wrapper function will also make it easier to
1364 adapt to future enhancements in your database driver. Note that at the
1365 time of this writing (October 2002), the DBI has no standardized way
1366 to deal with UTF-8 data. Please check the documentation to verify if
1370 my($self, $sth, $what) = @_; # $what is one of fetchrow_{array,hashref}
1376 my @arr = $sth->$what;
1378 defined && /[^\000-\177]/ && Encode::_utf8_on($_);
1382 my $ret = $sth->$what;
1384 for my $k (keys %$ret) {
1385 defined && /[^\000-\177]/ && Encode::_utf8_on($_) for $ret->{$k};
1389 defined && /[^\000-\177]/ && Encode::_utf8_on($_) for $ret;
1399 A large scalar that you know can only contain ASCII
1401 Scalars that contain only ASCII and are marked as UTF-8 are sometimes
1402 a drag to your program. If you recognize such a situation, just remove
1405 utf8::downgrade($val) if $] > 5.007;
1411 L<perluniintro>, L<encoding>, L<Encode>, L<open>, L<utf8>, L<bytes>,
1412 L<perlretut>, L<perlvar/"${^UNICODE}">
https://perl5.git.perl.org / perl5.git / blob