xref: /onnv-gate/usr/src/cmd/perl/5.8.4/distrib/pod/perlunicode.pod (revision 0:68f95e015346)
1=head1 NAME
2
3perlunicode - Unicode support in Perl
4
5=head1 DESCRIPTION
6
7=head2 Important Caveats
8
9Unicode support is an extensive requirement. While Perl does not
10implement the Unicode standard or the accompanying technical reports
11from cover to cover, Perl does support many Unicode features.
12
13=over 4
14
15=item Input and Output Layers
16
17Perl 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
19the ":utf8" layer.  Other encodings can be converted to Perl's
20encoding on input or from Perl's encoding on output by use of the
21":encoding(...)"  layer.  See L<open>.
22
23To indicate that Perl source itself is using a particular encoding,
24see L<encoding>.
25
26=item Regular Expressions
27
28The regular expression compiler produces polymorphic opcodes.  That is,
29the pattern adapts to the data and automatically switches to the Unicode
30character scheme when presented with Unicode data--or instead uses
31a traditional byte scheme when presented with byte data.
32
33=item C<use utf8> still needed to enable UTF-8/UTF-EBCDIC in scripts
34
35As a compatibility measure, the C<use utf8> pragma must be explicitly
36included to enable recognition of UTF-8 in the Perl scripts themselves
37(in string or regular expression literals, or in identifier names) on
38ASCII-based machines or to recognize UTF-EBCDIC on EBCDIC-based
39machines.  B<These are the only times when an explicit C<use utf8>
40is needed.>  See L<utf8>.
41
42You can also use the C<encoding> pragma to change the default encoding
43of the data in your script; see L<encoding>.
44
45=item BOM-marked scripts and UTF-16 scripts autodetected
46
47If a Perl script begins marked with the Unicode BOM (UTF-16LE, UTF16-BE,
48or UTF-8), or if the script looks like non-BOM-marked UTF-16 of either
49endianness, Perl will correctly read in the script as Unicode.
50(BOMless UTF-8 cannot be effectively recognized or differentiated from
51ISO 8859-1 or other eight-bit encodings.)
52
53=item C<use encoding> needed to upgrade non-Latin-1 byte strings
54
55By default, there is a fundamental asymmetry in Perl's unicode model:
56implicit upgrading from byte strings to Unicode strings assumes that
57they were encoded in I<ISO 8859-1 (Latin-1)>, but Unicode strings are
58downgraded with UTF-8 encoding.  This happens because the first 256
59codepoints in Unicode happens to agree with Latin-1.
60
61If you wish to interpret byte strings as UTF-8 instead, use the
62C<encoding> pragma:
63
64    use encoding 'utf8';
65
66See L</"Byte and Character Semantics"> for more details.
67
68=back
69
70=head2 Byte and Character Semantics
71
72Beginning with version 5.6, Perl uses logically-wide characters to
73represent strings internally.
74
75In future, Perl-level operations will be expected to work with
76characters rather than bytes.
77
78However, as an interim compatibility measure, Perl aims to
79provide a safe migration path from byte semantics to character
80semantics for programs.  For operations where Perl can unambiguously
81decide that the input data are characters, Perl switches to
82character semantics.  For operations where this determination cannot
83be made without additional information from the user, Perl decides in
84favor of compatibility and chooses to use byte semantics.
85
86This behavior preserves compatibility with earlier versions of Perl,
87which allowed byte semantics in Perl operations only if
88none of the program's inputs were marked as being as source of Unicode
89character data.  Such data may come from filehandles, from calls to
90external programs, from information provided by the system (such as %ENV),
91or from literals and constants in the source text.
92
93The C<bytes> pragma will always, regardless of platform, force byte
94semantics in a particular lexical scope.  See L<bytes>.
95
96The C<utf8> pragma is primarily a compatibility device that enables
97recognition of UTF-(8|EBCDIC) in literals encountered by the parser.
98Note that this pragma is only required while Perl defaults to byte
99semantics; when character semantics become the default, this pragma
100may become a no-op.  See L<utf8>.
101
102Unless explicitly stated, Perl operators use character semantics
103for Unicode data and byte semantics for non-Unicode data.
104The decision to use character semantics is made transparently.  If
105input data comes from a Unicode source--for example, if a character
106encoding layer is added to a filehandle or a literal Unicode
107string constant appears in a program--character semantics apply.
108Otherwise, byte semantics are in effect.  The C<bytes> pragma should
109be used to force byte semantics on Unicode data.
110
111If strings operating under byte semantics and strings with Unicode
112character data are concatenated, the new string will be created by
113decoding the byte strings as I<ISO 8859-1 (Latin-1)>, even if the
114old Unicode string used EBCDIC.  This translation is done without
115regard to the system's native 8-bit encoding.  To change this for
116systems with non-Latin-1 and non-EBCDIC native encodings, use the
117C<encoding> pragma.  See L<encoding>.
118
119Under character semantics, many operations that formerly operated on
120bytes now operate on characters. A character in Perl is
121logically just a number ranging from 0 to 2**31 or so. Larger
122characters may encode into longer sequences of bytes internally, but
123this internal detail is mostly hidden for Perl code.
124See L<perluniintro> for more.
125
126=head2 Effects of Character Semantics
127
128Character semantics have the following effects:
129
130=over 4
131
132=item *
133
134Strings--including hash keys--and regular expression patterns may
135contain characters that have an ordinal value larger than 255.
136
137If you use a Unicode editor to edit your program, Unicode characters
138may occur directly within the literal strings in one of the various
139Unicode encodings (UTF-8, UTF-EBCDIC, UCS-2, etc.), but will be recognized
140as such and converted to Perl's internal representation only if the
141appropriate L<encoding> is specified.
142
143Unicode characters can also be added to a string by using the
144C<\x{...}> notation.  The Unicode code for the desired character, in
145hexadecimal, should be placed in the braces. For instance, a smiley
146face is C<\x{263A}>.  This encoding scheme only works for characters
147with a code of 0x100 or above.
148
149Additionally, if you
150
151   use charnames ':full';
152
153you can use the C<\N{...}> notation and put the official Unicode
154character name within the braces, such as C<\N{WHITE SMILING FACE}>.
155
156
157=item *
158
159If an appropriate L<encoding> is specified, identifiers within the
160Perl script may contain Unicode alphanumeric characters, including
161ideographs.  Perl does not currently attempt to canonicalize variable
162names.
163
164=item *
165
166Regular expressions match characters instead of bytes.  "." matches
167a character instead of a byte.  The C<\C> pattern is provided to force
168a match a single byte--a C<char> in C, hence C<\C>.
169
170=item *
171
172Character classes in regular expressions match characters instead of
173bytes and match against the character properties specified in the
174Unicode properties database.  C<\w> can be used to match a Japanese
175ideograph, for instance.
176
177(However, and as a limitation of the current implementation, using
178C<\w> or C<\W> I<inside> a C<[...]> character class will still match
179with byte semantics.)
180
181=item *
182
183Named Unicode properties, scripts, and block ranges may be used like
184character classes via the C<\p{}> "matches property" construct and
185the  C<\P{}> negation, "doesn't match property".
186
187For instance, C<\p{Lu}> matches any character with the Unicode "Lu"
188(Letter, uppercase) property, while C<\p{M}> matches any character
189with an "M" (mark--accents and such) property.  Brackets are not
190required for single letter properties, so C<\p{M}> is equivalent to
191C<\pM>. Many predefined properties are available, such as
192C<\p{Mirrored}> and C<\p{Tibetan}>.
193
194The official Unicode script and block names have spaces and dashes as
195separators, but for convenience you can use dashes, spaces, or
196underbars, and case is unimportant. It is recommended, however, that
197for consistency you use the following naming: the official Unicode
198script, property, or block name (see below for the additional rules
199that apply to block names) with whitespace and dashes removed, and the
200words "uppercase-first-lowercase-rest". C<Latin-1 Supplement> thus
201becomes C<Latin1Supplement>.
202
203You can also use negation in both C<\p{}> and C<\P{}> by introducing a caret
204(^) between the first brace and the property name: C<\p{^Tamil}> is
205equal to C<\P{Tamil}>.
206
207B<NOTE: the properties, scripts, and blocks listed here are as of
208Unicode 3.2.0, March 2002, or Perl 5.8.0, July 2002.  Unicode 4.0.0
209came out in April 2003, and Perl 5.8.1 in September 2003.>
210
211Here are the basic Unicode General Category properties, followed by their
212long form.  You can use either; C<\p{Lu}> and C<\p{UppercaseLetter}>,
213for instance, are identical.
214
215    Short       Long
216
217    L           Letter
218    Lu          UppercaseLetter
219    Ll          LowercaseLetter
220    Lt          TitlecaseLetter
221    Lm          ModifierLetter
222    Lo          OtherLetter
223
224    M           Mark
225    Mn          NonspacingMark
226    Mc          SpacingMark
227    Me          EnclosingMark
228
229    N           Number
230    Nd          DecimalNumber
231    Nl          LetterNumber
232    No          OtherNumber
233
234    P           Punctuation
235    Pc          ConnectorPunctuation
236    Pd          DashPunctuation
237    Ps          OpenPunctuation
238    Pe          ClosePunctuation
239    Pi          InitialPunctuation
240                (may behave like Ps or Pe depending on usage)
241    Pf          FinalPunctuation
242                (may behave like Ps or Pe depending on usage)
243    Po          OtherPunctuation
244
245    S           Symbol
246    Sm          MathSymbol
247    Sc          CurrencySymbol
248    Sk          ModifierSymbol
249    So          OtherSymbol
250
251    Z           Separator
252    Zs          SpaceSeparator
253    Zl          LineSeparator
254    Zp          ParagraphSeparator
255
256    C           Other
257    Cc          Control
258    Cf          Format
259    Cs          Surrogate   (not usable)
260    Co          PrivateUse
261    Cn          Unassigned
262
263Single-letter properties match all characters in any of the
264two-letter sub-properties starting with the same letter.
265C<L&> is a special case, which is an alias for C<Ll>, C<Lu>, and C<Lt>.
266
267Because Perl hides the need for the user to understand the internal
268representation of Unicode characters, there is no need to implement
269the somewhat messy concept of surrogates. C<Cs> is therefore not
270supported.
271
272Because scripts differ in their directionality--Hebrew is
273written right to left, for example--Unicode supplies these properties:
274
275    Property    Meaning
276
277    BidiL       Left-to-Right
278    BidiLRE     Left-to-Right Embedding
279    BidiLRO     Left-to-Right Override
280    BidiR       Right-to-Left
281    BidiAL      Right-to-Left Arabic
282    BidiRLE     Right-to-Left Embedding
283    BidiRLO     Right-to-Left Override
284    BidiPDF     Pop Directional Format
285    BidiEN      European Number
286    BidiES      European Number Separator
287    BidiET      European Number Terminator
288    BidiAN      Arabic Number
289    BidiCS      Common Number Separator
290    BidiNSM     Non-Spacing Mark
291    BidiBN      Boundary Neutral
292    BidiB       Paragraph Separator
293    BidiS       Segment Separator
294    BidiWS      Whitespace
295    BidiON      Other Neutrals
296
297For example, C<\p{BidiR}> matches characters that are normally
298written right to left.
299
300=back
301
302=head2 Scripts
303
304The script names which can be used by C<\p{...}> and C<\P{...}>,
305such as in C<\p{Latin}> or C<\p{Cyrillic}>, are as follows:
306
307    Arabic
308    Armenian
309    Bengali
310    Bopomofo
311    Buhid
312    CanadianAboriginal
313    Cherokee
314    Cyrillic
315    Deseret
316    Devanagari
317    Ethiopic
318    Georgian
319    Gothic
320    Greek
321    Gujarati
322    Gurmukhi
323    Han
324    Hangul
325    Hanunoo
326    Hebrew
327    Hiragana
328    Inherited
329    Kannada
330    Katakana
331    Khmer
332    Lao
333    Latin
334    Malayalam
335    Mongolian
336    Myanmar
337    Ogham
338    OldItalic
339    Oriya
340    Runic
341    Sinhala
342    Syriac
343    Tagalog
344    Tagbanwa
345    Tamil
346    Telugu
347    Thaana
348    Thai
349    Tibetan
350    Yi
351
352Extended property classes can supplement the basic
353properties, defined by the F<PropList> Unicode database:
354
355    ASCIIHexDigit
356    BidiControl
357    Dash
358    Deprecated
359    Diacritic
360    Extender
361    GraphemeLink
362    HexDigit
363    Hyphen
364    Ideographic
365    IDSBinaryOperator
366    IDSTrinaryOperator
367    JoinControl
368    LogicalOrderException
369    NoncharacterCodePoint
370    OtherAlphabetic
371    OtherDefaultIgnorableCodePoint
372    OtherGraphemeExtend
373    OtherLowercase
374    OtherMath
375    OtherUppercase
376    QuotationMark
377    Radical
378    SoftDotted
379    TerminalPunctuation
380    UnifiedIdeograph
381    WhiteSpace
382
383and there are further derived properties:
384
385    Alphabetic      Lu + Ll + Lt + Lm + Lo + OtherAlphabetic
386    Lowercase       Ll + OtherLowercase
387    Uppercase       Lu + OtherUppercase
388    Math            Sm + OtherMath
389
390    ID_Start        Lu + Ll + Lt + Lm + Lo + Nl
391    ID_Continue     ID_Start + Mn + Mc + Nd + Pc
392
393    Any             Any character
394    Assigned        Any non-Cn character (i.e. synonym for \P{Cn})
395    Unassigned      Synonym for \p{Cn}
396    Common          Any character (or unassigned code point)
397                    not explicitly assigned to a script
398
399For backward compatibility (with Perl 5.6), all properties mentioned
400so far may have C<Is> prepended to their name, so C<\P{IsLu}>, for
401example, is equal to C<\P{Lu}>.
402
403=head2 Blocks
404
405In addition to B<scripts>, Unicode also defines B<blocks> of
406characters.  The difference between scripts and blocks is that the
407concept of scripts is closer to natural languages, while the concept
408of blocks is more of an artificial grouping based on groups of 256
409Unicode characters. For example, the C<Latin> script contains letters
410from many blocks but does not contain all the characters from those
411blocks. It does not, for example, contain digits, because digits are
412shared across many scripts. Digits and similar groups, like
413punctuation, are in a category called C<Common>.
414
415For more about scripts, see the UTR #24:
416
417   http://www.unicode.org/unicode/reports/tr24/
418
419For more about blocks, see:
420
421   http://www.unicode.org/Public/UNIDATA/Blocks.txt
422
423Block names are given with the C<In> prefix. For example, the
424Katakana block is referenced via C<\p{InKatakana}>.  The C<In>
425prefix may be omitted if there is no naming conflict with a script
426or any other property, but it is recommended that C<In> always be used
427for block tests to avoid confusion.
428
429These block names are supported:
430
431    InAlphabeticPresentationForms
432    InArabic
433    InArabicPresentationFormsA
434    InArabicPresentationFormsB
435    InArmenian
436    InArrows
437    InBasicLatin
438    InBengali
439    InBlockElements
440    InBopomofo
441    InBopomofoExtended
442    InBoxDrawing
443    InBraillePatterns
444    InBuhid
445    InByzantineMusicalSymbols
446    InCJKCompatibility
447    InCJKCompatibilityForms
448    InCJKCompatibilityIdeographs
449    InCJKCompatibilityIdeographsSupplement
450    InCJKRadicalsSupplement
451    InCJKSymbolsAndPunctuation
452    InCJKUnifiedIdeographs
453    InCJKUnifiedIdeographsExtensionA
454    InCJKUnifiedIdeographsExtensionB
455    InCherokee
456    InCombiningDiacriticalMarks
457    InCombiningDiacriticalMarksforSymbols
458    InCombiningHalfMarks
459    InControlPictures
460    InCurrencySymbols
461    InCyrillic
462    InCyrillicSupplementary
463    InDeseret
464    InDevanagari
465    InDingbats
466    InEnclosedAlphanumerics
467    InEnclosedCJKLettersAndMonths
468    InEthiopic
469    InGeneralPunctuation
470    InGeometricShapes
471    InGeorgian
472    InGothic
473    InGreekExtended
474    InGreekAndCoptic
475    InGujarati
476    InGurmukhi
477    InHalfwidthAndFullwidthForms
478    InHangulCompatibilityJamo
479    InHangulJamo
480    InHangulSyllables
481    InHanunoo
482    InHebrew
483    InHighPrivateUseSurrogates
484    InHighSurrogates
485    InHiragana
486    InIPAExtensions
487    InIdeographicDescriptionCharacters
488    InKanbun
489    InKangxiRadicals
490    InKannada
491    InKatakana
492    InKatakanaPhoneticExtensions
493    InKhmer
494    InLao
495    InLatin1Supplement
496    InLatinExtendedA
497    InLatinExtendedAdditional
498    InLatinExtendedB
499    InLetterlikeSymbols
500    InLowSurrogates
501    InMalayalam
502    InMathematicalAlphanumericSymbols
503    InMathematicalOperators
504    InMiscellaneousMathematicalSymbolsA
505    InMiscellaneousMathematicalSymbolsB
506    InMiscellaneousSymbols
507    InMiscellaneousTechnical
508    InMongolian
509    InMusicalSymbols
510    InMyanmar
511    InNumberForms
512    InOgham
513    InOldItalic
514    InOpticalCharacterRecognition
515    InOriya
516    InPrivateUseArea
517    InRunic
518    InSinhala
519    InSmallFormVariants
520    InSpacingModifierLetters
521    InSpecials
522    InSuperscriptsAndSubscripts
523    InSupplementalArrowsA
524    InSupplementalArrowsB
525    InSupplementalMathematicalOperators
526    InSupplementaryPrivateUseAreaA
527    InSupplementaryPrivateUseAreaB
528    InSyriac
529    InTagalog
530    InTagbanwa
531    InTags
532    InTamil
533    InTelugu
534    InThaana
535    InThai
536    InTibetan
537    InUnifiedCanadianAboriginalSyllabics
538    InVariationSelectors
539    InYiRadicals
540    InYiSyllables
541
542=over 4
543
544=item *
545
546The special pattern C<\X> matches any extended Unicode
547sequence--"a combining character sequence" in Standardese--where the
548first character is a base character and subsequent characters are mark
549characters that apply to the base character.  C<\X> is equivalent to
550C<(?:\PM\pM*)>.
551
552=item *
553
554The C<tr///> operator translates characters instead of bytes.  Note
555that the C<tr///CU> functionality has been removed.  For similar
556functionality see pack('U0', ...) and pack('C0', ...).
557
558=item *
559
560Case translation operators use the Unicode case translation tables
561when character input is provided.  Note that C<uc()>, or C<\U> in
562interpolated strings, translates to uppercase, while C<ucfirst>,
563or C<\u> in interpolated strings, translates to titlecase in languages
564that make the distinction.
565
566=item *
567
568Most operators that deal with positions or lengths in a string will
569automatically switch to using character positions, including
570C<chop()>, C<chomp()>, C<substr()>, C<pos()>, C<index()>, C<rindex()>,
571C<sprintf()>, C<write()>, and C<length()>.  Operators that
572specifically do not switch include C<vec()>, C<pack()>, and
573C<unpack()>.  Operators that really don't care include
574operators that treats strings as a bucket of bits such as C<sort()>,
575and operators dealing with filenames.
576
577=item *
578
579The C<pack()>/C<unpack()> letters C<c> and C<C> do I<not> change,
580since they are often used for byte-oriented formats.  Again, think
581C<char> in the C language.
582
583There is a new C<U> specifier that converts between Unicode characters
584and code points.
585
586=item *
587
588The C<chr()> and C<ord()> functions work on characters, similar to
589C<pack("U")> and C<unpack("U")>, I<not> C<pack("C")> and
590C<unpack("C")>.  C<pack("C")> and C<unpack("C")> are methods for
591emulating byte-oriented C<chr()> and C<ord()> on Unicode strings.
592While these methods reveal the internal encoding of Unicode strings,
593that is not something one normally needs to care about at all.
594
595=item *
596
597The bit string operators, C<& | ^ ~>, can operate on character data.
598However, for backward compatibility, such as when using bit string
599operations when characters are all less than 256 in ordinal value, one
600should not use C<~> (the bit complement) with characters of both
601values less than 256 and values greater than 256.  Most importantly,
602DeMorgan's laws (C<~($x|$y) eq ~$x&~$y> and C<~($x&$y) eq ~$x|~$y>)
603will not hold.  The reason for this mathematical I<faux pas> is that
604the complement cannot return B<both> the 8-bit (byte-wide) bit
605complement B<and> the full character-wide bit complement.
606
607=item *
608
609lc(), uc(), lcfirst(), and ucfirst() work for the following cases:
610
611=over 8
612
613=item *
614
615the case mapping is from a single Unicode character to another
616single Unicode character, or
617
618=item *
619
620the case mapping is from a single Unicode character to more
621than one Unicode character.
622
623=back
624
625Things to do with locales (Lithuanian, Turkish, Azeri) do B<not> work
626since Perl does not understand the concept of Unicode locales.
627
628See the Unicode Technical Report #21, Case Mappings, for more details.
629
630=back
631
632=over 4
633
634=item *
635
636And finally, C<scalar reverse()> reverses by character rather than by byte.
637
638=back
639
640=head2 User-Defined Character Properties
641
642You can define your own character properties by defining subroutines
643whose names begin with "In" or "Is".  The subroutines must be defined
644in the C<main> package.  The user-defined properties can be used in the
645regular expression C<\p> and C<\P> constructs.  Note that the effect
646is compile-time and immutable once defined.
647
648The subroutines must return a specially-formatted string, with one
649or more newline-separated lines.  Each line must be one of the following:
650
651=over 4
652
653=item *
654
655Two hexadecimal numbers separated by horizontal whitespace (space or
656tabular characters) denoting a range of Unicode code points to include.
657
658=item *
659
660Something to include, prefixed by "+": a built-in character
661property (prefixed by "utf8::"), to represent all the characters in that
662property; two hexadecimal code points for a range; or a single
663hexadecimal code point.
664
665=item *
666
667Something to exclude, prefixed by "-": an existing character
668property (prefixed by "utf8::"), for all the characters in that
669property; two hexadecimal code points for a range; or a single
670hexadecimal code point.
671
672=item *
673
674Something to negate, prefixed "!": an existing character
675property (prefixed by "utf8::") for all the characters except the
676characters in the property; two hexadecimal code points for a range;
677or a single hexadecimal code point.
678
679=back
680
681For example, to define a property that covers both the Japanese
682syllabaries (hiragana and katakana), you can define
683
684    sub InKana {
685	return <<END;
686    3040\t309F
687    30A0\t30FF
688    END
689    }
690
691Imagine that the here-doc end marker is at the beginning of the line.
692Now you can use C<\p{InKana}> and C<\P{InKana}>.
693
694You could also have used the existing block property names:
695
696    sub InKana {
697	return <<'END';
698    +utf8::InHiragana
699    +utf8::InKatakana
700    END
701    }
702
703Suppose you wanted to match only the allocated characters,
704not the raw block ranges: in other words, you want to remove
705the non-characters:
706
707    sub InKana {
708	return <<'END';
709    +utf8::InHiragana
710    +utf8::InKatakana
711    -utf8::IsCn
712    END
713    }
714
715The negation is useful for defining (surprise!) negated classes.
716
717    sub InNotKana {
718	return <<'END';
719    !utf8::InHiragana
720    -utf8::InKatakana
721    +utf8::IsCn
722    END
723    }
724
725You can also define your own mappings to be used in the lc(),
726lcfirst(), uc(), and ucfirst() (or their string-inlined versions).
727The principle is the same: define subroutines in the C<main> package
728with names like C<ToLower> (for lc() and lcfirst()), C<ToTitle> (for
729the first character in ucfirst()), and C<ToUpper> (for uc(), and the
730rest of the characters in ucfirst()).
731
732The string returned by the subroutines needs now to be three
733hexadecimal numbers separated by tabulators: start of the source
734range, end of the source range, and start of the destination range.
735For example:
736
737    sub ToUpper {
738	return <<END;
739    0061\t0063\t0041
740    END
741    }
742
743defines an uc() mapping that causes only the characters "a", "b", and
744"c" to be mapped to "A", "B", "C", all other characters will remain
745unchanged.
746
747If there is no source range to speak of, that is, the mapping is from
748a single character to another single character, leave the end of the
749source range empty, but the two tabulator characters are still needed.
750For example:
751
752    sub ToLower {
753	return <<END;
754    0041\t\t0061
755    END
756    }
757
758defines a lc() mapping that causes only "A" to be mapped to "a", all
759other characters will remain unchanged.
760
761(For serious hackers only)  If you want to introspect the default
762mappings, you can find the data in the directory
763C<$Config{privlib}>/F<unicore/To/>.  The mapping data is returned as
764the here-document, and the C<utf8::ToSpecFoo> are special exception
765mappings derived from <$Config{privlib}>/F<unicore/SpecialCasing.txt>.
766The C<Digit> and C<Fold> mappings that one can see in the directory
767are not directly user-accessible, one can use either the
768C<Unicode::UCD> module, or just match case-insensitively (that's when
769the C<Fold> mapping is used).
770
771A final note on the user-defined property tests and mappings: they
772will be used only if the scalar has been marked as having Unicode
773characters.  Old byte-style strings will not be affected.
774
775=head2 Character Encodings for Input and Output
776
777See L<Encode>.
778
779=head2 Unicode Regular Expression Support Level
780
781The following list of Unicode support for regular expressions describes
782all the features currently supported.  The references to "Level N"
783and the section numbers refer to the Unicode Technical Report 18,
784"Unicode Regular Expression Guidelines", version 6 (Unicode 3.2.0,
785Perl 5.8.0).
786
787=over 4
788
789=item *
790
791Level 1 - Basic Unicode Support
792
793        2.1 Hex Notation                        - done          [1]
794            Named Notation                      - done          [2]
795        2.2 Categories                          - done          [3][4]
796        2.3 Subtraction                         - MISSING       [5][6]
797        2.4 Simple Word Boundaries              - done          [7]
798        2.5 Simple Loose Matches                - done          [8]
799        2.6 End of Line                         - MISSING       [9][10]
800
801        [ 1] \x{...}
802        [ 2] \N{...}
803        [ 3] . \p{...} \P{...}
804        [ 4] now scripts (see UTR#24 Script Names) in addition to blocks
805        [ 5] have negation
806        [ 6] can use regular expression look-ahead [a]
807             or user-defined character properties [b] to emulate subtraction
808        [ 7] include Letters in word characters
809        [ 8] note that Perl does Full case-folding in matching, not Simple:
810             for example U+1F88 is equivalent with U+1F00 U+03B9,
811             not with 1F80.  This difference matters for certain Greek
812             capital letters with certain modifiers: the Full case-folding
813             decomposes the letter, while the Simple case-folding would map
814             it to a single character.
815        [ 9] see UTR #13 Unicode Newline Guidelines
816        [10] should do ^ and $ also on \x{85}, \x{2028} and \x{2029}
817             (should also affect <>, $., and script line numbers)
818             (the \x{85}, \x{2028} and \x{2029} do match \s)
819
820[a] You can mimic class subtraction using lookahead.
821For example, what UTR #18 might write as
822
823    [{Greek}-[{UNASSIGNED}]]
824
825in Perl can be written as:
826
827    (?!\p{Unassigned})\p{InGreekAndCoptic}
828    (?=\p{Assigned})\p{InGreekAndCoptic}
829
830But in this particular example, you probably really want
831
832    \p{GreekAndCoptic}
833
834which will match assigned characters known to be part of the Greek script.
835
836Also see the Unicode::Regex::Set module, it does implement the full
837UTR #18 grouping, intersection, union, and removal (subtraction) syntax.
838
839[b] See L</"User-Defined Character Properties">.
840
841=item *
842
843Level 2 - Extended Unicode Support
844
845        3.1 Surrogates                          - MISSING	[11]
846        3.2 Canonical Equivalents               - MISSING       [12][13]
847        3.3 Locale-Independent Graphemes        - MISSING       [14]
848        3.4 Locale-Independent Words            - MISSING       [15]
849        3.5 Locale-Independent Loose Matches    - MISSING       [16]
850
851        [11] Surrogates are solely a UTF-16 concept and Perl's internal
852             representation is UTF-8.  The Encode module does UTF-16, though.
853        [12] see UTR#15 Unicode Normalization
854        [13] have Unicode::Normalize but not integrated to regexes
855        [14] have \X but at this level . should equal that
856        [15] need three classes, not just \w and \W
857        [16] see UTR#21 Case Mappings
858
859=item *
860
861Level 3 - Locale-Sensitive Support
862
863        4.1 Locale-Dependent Categories         - MISSING
864        4.2 Locale-Dependent Graphemes          - MISSING       [16][17]
865        4.3 Locale-Dependent Words              - MISSING
866        4.4 Locale-Dependent Loose Matches      - MISSING
867        4.5 Locale-Dependent Ranges             - MISSING
868
869        [16] see UTR#10 Unicode Collation Algorithms
870        [17] have Unicode::Collate but not integrated to regexes
871
872=back
873
874=head2 Unicode Encodings
875
876Unicode characters are assigned to I<code points>, which are abstract
877numbers.  To use these numbers, various encodings are needed.
878
879=over 4
880
881=item *
882
883UTF-8
884
885UTF-8 is a variable-length (1 to 6 bytes, current character allocations
886require 4 bytes), byte-order independent encoding. For ASCII (and we
887really do mean 7-bit ASCII, not another 8-bit encoding), UTF-8 is
888transparent.
889
890The following table is from Unicode 3.2.
891
892 Code Points            1st Byte  2nd Byte  3rd Byte  4th Byte
893
894   U+0000..U+007F       00..7F
895   U+0080..U+07FF       C2..DF    80..BF
896   U+0800..U+0FFF       E0        A0..BF    80..BF
897   U+1000..U+CFFF       E1..EC    80..BF    80..BF
898   U+D000..U+D7FF       ED        80..9F    80..BF
899   U+D800..U+DFFF       ******* ill-formed *******
900   U+E000..U+FFFF       EE..EF    80..BF    80..BF
901  U+10000..U+3FFFF      F0        90..BF    80..BF    80..BF
902  U+40000..U+FFFFF      F1..F3    80..BF    80..BF    80..BF
903 U+100000..U+10FFFF     F4        80..8F    80..BF    80..BF
904
905Note the C<A0..BF> in C<U+0800..U+0FFF>, the C<80..9F> in
906C<U+D000...U+D7FF>, the C<90..B>F in C<U+10000..U+3FFFF>, and the
907C<80...8F> in C<U+100000..U+10FFFF>.  The "gaps" are caused by legal
908UTF-8 avoiding non-shortest encodings: it is technically possible to
909UTF-8-encode a single code point in different ways, but that is
910explicitly forbidden, and the shortest possible encoding should always
911be used.  So that's what Perl does.
912
913Another way to look at it is via bits:
914
915 Code Points                    1st Byte   2nd Byte  3rd Byte  4th Byte
916
917                    0aaaaaaa     0aaaaaaa
918            00000bbbbbaaaaaa     110bbbbb  10aaaaaa
919            ccccbbbbbbaaaaaa     1110cccc  10bbbbbb  10aaaaaa
920  00000dddccccccbbbbbbaaaaaa     11110ddd  10cccccc  10bbbbbb  10aaaaaa
921
922As you can see, the continuation bytes all begin with C<10>, and the
923leading bits of the start byte tell how many bytes the are in the
924encoded character.
925
926=item *
927
928UTF-EBCDIC
929
930Like UTF-8 but EBCDIC-safe, in the way that UTF-8 is ASCII-safe.
931
932=item *
933
934UTF-16, UTF-16BE, UTF-16LE, Surrogates, and BOMs (Byte Order Marks)
935
936The followings items are mostly for reference and general Unicode
937knowledge, Perl doesn't use these constructs internally.
938
939UTF-16 is a 2 or 4 byte encoding.  The Unicode code points
940C<U+0000..U+FFFF> are stored in a single 16-bit unit, and the code
941points C<U+10000..U+10FFFF> in two 16-bit units.  The latter case is
942using I<surrogates>, the first 16-bit unit being the I<high
943surrogate>, and the second being the I<low surrogate>.
944
945Surrogates are code points set aside to encode the C<U+10000..U+10FFFF>
946range of Unicode code points in pairs of 16-bit units.  The I<high
947surrogates> are the range C<U+D800..U+DBFF>, and the I<low surrogates>
948are the range C<U+DC00..U+DFFF>.  The surrogate encoding is
949
950	$hi = ($uni - 0x10000) / 0x400 + 0xD800;
951	$lo = ($uni - 0x10000) % 0x400 + 0xDC00;
952
953and the decoding is
954
955	$uni = 0x10000 + ($hi - 0xD800) * 0x400 + ($lo - 0xDC00);
956
957If you try to generate surrogates (for example by using chr()), you
958will get a warning if warnings are turned on, because those code
959points are not valid for a Unicode character.
960
961Because of the 16-bitness, UTF-16 is byte-order dependent.  UTF-16
962itself can be used for in-memory computations, but if storage or
963transfer is required either UTF-16BE (big-endian) or UTF-16LE
964(little-endian) encodings must be chosen.
965
966This introduces another problem: what if you just know that your data
967is UTF-16, but you don't know which endianness?  Byte Order Marks, or
968BOMs, are a solution to this.  A special character has been reserved
969in Unicode to function as a byte order marker: the character with the
970code point C<U+FEFF> is the BOM.
971
972The trick is that if you read a BOM, you will know the byte order,
973since if it was written on a big-endian platform, you will read the
974bytes C<0xFE 0xFF>, but if it was written on a little-endian platform,
975you will read the bytes C<0xFF 0xFE>.  (And if the originating platform
976was writing in UTF-8, you will read the bytes C<0xEF 0xBB 0xBF>.)
977
978The way this trick works is that the character with the code point
979C<U+FFFE> is guaranteed not to be a valid Unicode character, so the
980sequence of bytes C<0xFF 0xFE> is unambiguously "BOM, represented in
981little-endian format" and cannot be C<U+FFFE>, represented in big-endian
982format".
983
984=item *
985
986UTF-32, UTF-32BE, UTF-32LE
987
988The UTF-32 family is pretty much like the UTF-16 family, expect that
989the units are 32-bit, and therefore the surrogate scheme is not
990needed.  The BOM signatures will be C<0x00 0x00 0xFE 0xFF> for BE and
991C<0xFF 0xFE 0x00 0x00> for LE.
992
993=item *
994
995UCS-2, UCS-4
996
997Encodings defined by the ISO 10646 standard.  UCS-2 is a 16-bit
998encoding.  Unlike UTF-16, UCS-2 is not extensible beyond C<U+FFFF>,
999because it does not use surrogates.  UCS-4 is a 32-bit encoding,
1000functionally identical to UTF-32.
1001
1002=item *
1003
1004UTF-7
1005
1006A seven-bit safe (non-eight-bit) encoding, which is useful if the
1007transport or storage is not eight-bit safe.  Defined by RFC 2152.
1008
1009=back
1010
1011=head2 Security Implications of Unicode
1012
1013=over 4
1014
1015=item *
1016
1017Malformed UTF-8
1018
1019Unfortunately, the specification of UTF-8 leaves some room for
1020interpretation of how many bytes of encoded output one should generate
1021from one input Unicode character.  Strictly speaking, the shortest
1022possible sequence of UTF-8 bytes should be generated,
1023because otherwise there is potential for an input buffer overflow at
1024the receiving end of a UTF-8 connection.  Perl always generates the
1025shortest length UTF-8, and with warnings on Perl will warn about
1026non-shortest length UTF-8 along with other malformations, such as the
1027surrogates, which are not real Unicode code points.
1028
1029=item *
1030
1031Regular expressions behave slightly differently between byte data and
1032character (Unicode) data.  For example, the "word character" character
1033class C<\w> will work differently depending on if data is eight-bit bytes
1034or Unicode.
1035
1036In the first case, the set of C<\w> characters is either small--the
1037default set of alphabetic characters, digits, and the "_"--or, if you
1038are using a locale (see L<perllocale>), the C<\w> might contain a few
1039more letters according to your language and country.
1040
1041In the second case, the C<\w> set of characters is much, much larger.
1042Most importantly, even in the set of the first 256 characters, it will
1043probably match different characters: unlike most locales, which are
1044specific to a language and country pair, Unicode classifies all the
1045characters that are letters I<somewhere> as C<\w>.  For example, your
1046locale might not think that LATIN SMALL LETTER ETH is a letter (unless
1047you happen to speak Icelandic), but Unicode does.
1048
1049As discussed elsewhere, Perl has one foot (two hooves?) planted in
1050each of two worlds: the old world of bytes and the new world of
1051characters, upgrading from bytes to characters when necessary.
1052If your legacy code does not explicitly use Unicode, no automatic
1053switch-over to characters should happen.  Characters shouldn't get
1054downgraded to bytes, either.  It is possible to accidentally mix bytes
1055and characters, however (see L<perluniintro>), in which case C<\w> in
1056regular expressions might start behaving differently.  Review your
1057code.  Use warnings and the C<strict> pragma.
1058
1059=back
1060
1061=head2 Unicode in Perl on EBCDIC
1062
1063The way Unicode is handled on EBCDIC platforms is still
1064experimental.  On such platforms, references to UTF-8 encoding in this
1065document and elsewhere should be read as meaning the UTF-EBCDIC
1066specified in Unicode Technical Report 16, unless ASCII vs. EBCDIC issues
1067are specifically discussed. There is no C<utfebcdic> pragma or
1068":utfebcdic" layer; rather, "utf8" and ":utf8" are reused to mean
1069the platform's "natural" 8-bit encoding of Unicode. See L<perlebcdic>
1070for more discussion of the issues.
1071
1072=head2 Locales
1073
1074Usually locale settings and Unicode do not affect each other, but
1075there are a couple of exceptions:
1076
1077=over 4
1078
1079=item *
1080
1081You can enable automatic UTF-8-ification of your standard file
1082handles, default C<open()> layer, and C<@ARGV> by using either
1083the C<-C> command line switch or the C<PERL_UNICODE> environment
1084variable, see L<perlrun> for the documentation of the C<-C> switch.
1085
1086=item *
1087
1088Perl tries really hard to work both with Unicode and the old
1089byte-oriented world. Most often this is nice, but sometimes Perl's
1090straddling of the proverbial fence causes problems.
1091
1092=back
1093
1094=head2 When Unicode Does Not Happen
1095
1096While Perl does have extensive ways to input and output in Unicode,
1097and few other 'entry points' like the @ARGV which can be interpreted
1098as Unicode (UTF-8), there still are many places where Unicode (in some
1099encoding or another) could be given as arguments or received as
1100results, or both, but it is not.
1101
1102The following are such interfaces.  For all of these interfaces Perl
1103currently (as of 5.8.3) simply assumes byte strings both as arguments
1104and results, or UTF-8 strings if the C<encoding> pragma has been used.
1105
1106One reason why Perl does not attempt to resolve the role of Unicode in
1107this cases is that the answers are highly dependent on the operating
1108system and the file system(s).  For example, whether filenames can be
1109in Unicode, and in exactly what kind of encoding, is not exactly a
1110portable concept.  Similarly for the qx and system: how well will the
1111'command line interface' (and which of them?) handle Unicode?
1112
1113=over 4
1114
1115=item *
1116
1117chmod, chmod, chown, chroot, exec, link, lstat, mkdir,
1118rename, rmdir, stat, symlink, truncate, unlink, utime, -X
1119
1120=item *
1121
1122%ENV
1123
1124=item *
1125
1126glob (aka the <*>)
1127
1128=item *
1129
1130open, opendir, sysopen
1131
1132=item *
1133
1134qx (aka the backtick operator), system
1135
1136=item *
1137
1138readdir, readlink
1139
1140=back
1141
1142=head2 Forcing Unicode in Perl (Or Unforcing Unicode in Perl)
1143
1144Sometimes (see L</"When Unicode Does Not Happen">) there are
1145situations where you simply need to force Perl to believe that a byte
1146string is UTF-8, or vice versa.  The low-level calls
1147utf8::upgrade($bytestring) and utf8::downgrade($utf8string) are
1148the answers.
1149
1150Do not use them without careful thought, though: Perl may easily get
1151very confused, angry, or even crash, if you suddenly change the 'nature'
1152of scalar like that.  Especially careful you have to be if you use the
1153utf8::upgrade(): any random byte string is not valid UTF-8.
1154
1155=head2 Using Unicode in XS
1156
1157If you want to handle Perl Unicode in XS extensions, you may find the
1158following C APIs useful.  See also L<perlguts/"Unicode Support"> for an
1159explanation about Unicode at the XS level, and L<perlapi> for the API
1160details.
1161
1162=over 4
1163
1164=item *
1165
1166C<DO_UTF8(sv)> returns true if the C<UTF8> flag is on and the bytes
1167pragma is not in effect.  C<SvUTF8(sv)> returns true is the C<UTF8>
1168flag is on; the bytes pragma is ignored.  The C<UTF8> flag being on
1169does B<not> mean that there are any characters of code points greater
1170than 255 (or 127) in the scalar or that there are even any characters
1171in the scalar.  What the C<UTF8> flag means is that the sequence of
1172octets in the representation of the scalar is the sequence of UTF-8
1173encoded code points of the characters of a string.  The C<UTF8> flag
1174being off means that each octet in this representation encodes a
1175single character with code point 0..255 within the string.  Perl's
1176Unicode model is not to use UTF-8 until it is absolutely necessary.
1177
1178=item *
1179
1180C<uvuni_to_utf8(buf, chr)> writes a Unicode character code point into
1181a buffer encoding the code point as UTF-8, and returns a pointer
1182pointing after the UTF-8 bytes.
1183
1184=item *
1185
1186C<utf8_to_uvuni(buf, lenp)> reads UTF-8 encoded bytes from a buffer and
1187returns the Unicode character code point and, optionally, the length of
1188the UTF-8 byte sequence.
1189
1190=item *
1191
1192C<utf8_length(start, end)> returns the length of the UTF-8 encoded buffer
1193in characters.  C<sv_len_utf8(sv)> returns the length of the UTF-8 encoded
1194scalar.
1195
1196=item *
1197
1198C<sv_utf8_upgrade(sv)> converts the string of the scalar to its UTF-8
1199encoded form.  C<sv_utf8_downgrade(sv)> does the opposite, if
1200possible.  C<sv_utf8_encode(sv)> is like sv_utf8_upgrade except that
1201it does not set the C<UTF8> flag.  C<sv_utf8_decode()> does the
1202opposite of C<sv_utf8_encode()>.  Note that none of these are to be
1203used as general-purpose encoding or decoding interfaces: C<use Encode>
1204for that.  C<sv_utf8_upgrade()> is affected by the encoding pragma
1205but C<sv_utf8_downgrade()> is not (since the encoding pragma is
1206designed to be a one-way street).
1207
1208=item *
1209
1210C<is_utf8_char(s)> returns true if the pointer points to a valid UTF-8
1211character.
1212
1213=item *
1214
1215C<is_utf8_string(buf, len)> returns true if C<len> bytes of the buffer
1216are valid UTF-8.
1217
1218=item *
1219
1220C<UTF8SKIP(buf)> will return the number of bytes in the UTF-8 encoded
1221character in the buffer.  C<UNISKIP(chr)> will return the number of bytes
1222required to UTF-8-encode the Unicode character code point.  C<UTF8SKIP()>
1223is useful for example for iterating over the characters of a UTF-8
1224encoded buffer; C<UNISKIP()> is useful, for example, in computing
1225the size required for a UTF-8 encoded buffer.
1226
1227=item *
1228
1229C<utf8_distance(a, b)> will tell the distance in characters between the
1230two pointers pointing to the same UTF-8 encoded buffer.
1231
1232=item *
1233
1234C<utf8_hop(s, off)> will return a pointer to an UTF-8 encoded buffer
1235that is C<off> (positive or negative) Unicode characters displaced
1236from the UTF-8 buffer C<s>.  Be careful not to overstep the buffer:
1237C<utf8_hop()> will merrily run off the end or the beginning of the
1238buffer if told to do so.
1239
1240=item *
1241
1242C<pv_uni_display(dsv, spv, len, pvlim, flags)> and
1243C<sv_uni_display(dsv, ssv, pvlim, flags)> are useful for debugging the
1244output of Unicode strings and scalars.  By default they are useful
1245only for debugging--they display B<all> characters as hexadecimal code
1246points--but with the flags C<UNI_DISPLAY_ISPRINT>,
1247C<UNI_DISPLAY_BACKSLASH>, and C<UNI_DISPLAY_QQ> you can make the
1248output more readable.
1249
1250=item *
1251
1252C<ibcmp_utf8(s1, pe1, u1, l1, u1, s2, pe2, l2, u2)> can be used to
1253compare two strings case-insensitively in Unicode.  For case-sensitive
1254comparisons you can just use C<memEQ()> and C<memNE()> as usual.
1255
1256=back
1257
1258For more information, see L<perlapi>, and F<utf8.c> and F<utf8.h>
1259in the Perl source code distribution.
1260
1261=head1 BUGS
1262
1263=head2 Interaction with Locales
1264
1265Use of locales with Unicode data may lead to odd results.  Currently,
1266Perl attempts to attach 8-bit locale info to characters in the range
12670..255, but this technique is demonstrably incorrect for locales that
1268use characters above that range when mapped into Unicode.  Perl's
1269Unicode support will also tend to run slower.  Use of locales with
1270Unicode is discouraged.
1271
1272=head2 Interaction with Extensions
1273
1274When Perl exchanges data with an extension, the extension should be
1275able to understand the UTF-8 flag and act accordingly. If the
1276extension doesn't know about the flag, it's likely that the extension
1277will return incorrectly-flagged data.
1278
1279So if you're working with Unicode data, consult the documentation of
1280every module you're using if there are any issues with Unicode data
1281exchange. If the documentation does not talk about Unicode at all,
1282suspect the worst and probably look at the source to learn how the
1283module is implemented. Modules written completely in Perl shouldn't
1284cause problems. Modules that directly or indirectly access code written
1285in other programming languages are at risk.
1286
1287For affected functions, the simple strategy to avoid data corruption is
1288to always make the encoding of the exchanged data explicit. Choose an
1289encoding that you know the extension can handle. Convert arguments passed
1290to the extensions to that encoding and convert results back from that
1291encoding. Write wrapper functions that do the conversions for you, so
1292you can later change the functions when the extension catches up.
1293
1294To provide an example, let's say the popular Foo::Bar::escape_html
1295function doesn't deal with Unicode data yet. The wrapper function
1296would convert the argument to raw UTF-8 and convert the result back to
1297Perl's internal representation like so:
1298
1299    sub my_escape_html ($) {
1300      my($what) = shift;
1301      return unless defined $what;
1302      Encode::decode_utf8(Foo::Bar::escape_html(Encode::encode_utf8($what)));
1303    }
1304
1305Sometimes, when the extension does not convert data but just stores
1306and retrieves them, you will be in a position to use the otherwise
1307dangerous Encode::_utf8_on() function. Let's say the popular
1308C<Foo::Bar> extension, written in C, provides a C<param> method that
1309lets you store and retrieve data according to these prototypes:
1310
1311    $self->param($name, $value);            # set a scalar
1312    $value = $self->param($name);           # retrieve a scalar
1313
1314If it does not yet provide support for any encoding, one could write a
1315derived class with such a C<param> method:
1316
1317    sub param {
1318      my($self,$name,$value) = @_;
1319      utf8::upgrade($name);     # make sure it is UTF-8 encoded
1320      if (defined $value)
1321        utf8::upgrade($value);  # make sure it is UTF-8 encoded
1322        return $self->SUPER::param($name,$value);
1323      } else {
1324        my $ret = $self->SUPER::param($name);
1325        Encode::_utf8_on($ret); # we know, it is UTF-8 encoded
1326        return $ret;
1327      }
1328    }
1329
1330Some extensions provide filters on data entry/exit points, such as
1331DB_File::filter_store_key and family. Look out for such filters in
1332the documentation of your extensions, they can make the transition to
1333Unicode data much easier.
1334
1335=head2 Speed
1336
1337Some functions are slower when working on UTF-8 encoded strings than
1338on byte encoded strings.  All functions that need to hop over
1339characters such as length(), substr() or index(), or matching regular
1340expressions can work B<much> faster when the underlying data are
1341byte-encoded.
1342
1343In Perl 5.8.0 the slowness was often quite spectacular; in Perl 5.8.1
1344a caching scheme was introduced which will hopefully make the slowness
1345somewhat less spectacular, at least for some operations.  In general,
1346operations with UTF-8 encoded strings are still slower. As an example,
1347the Unicode properties (character classes) like C<\p{Nd}> are known to
1348be quite a bit slower (5-20 times) than their simpler counterparts
1349like C<\d> (then again, there 268 Unicode characters matching C<Nd>
1350compared with the 10 ASCII characters matching C<d>).
1351
1352=head2 Porting code from perl-5.6.X
1353
1354Perl 5.8 has a different Unicode model from 5.6. In 5.6 the programmer
1355was required to use the C<utf8> pragma to declare that a given scope
1356expected to deal with Unicode data and had to make sure that only
1357Unicode data were reaching that scope. If you have code that is
1358working with 5.6, you will need some of the following adjustments to
1359your code. The examples are written such that the code will continue
1360to work under 5.6, so you should be safe to try them out.
1361
1362=over 4
1363
1364=item *
1365
1366A filehandle that should read or write UTF-8
1367
1368  if ($] > 5.007) {
1369    binmode $fh, ":utf8";
1370  }
1371
1372=item *
1373
1374A scalar that is going to be passed to some extension
1375
1376Be it Compress::Zlib, Apache::Request or any extension that has no
1377mention of Unicode in the manpage, you need to make sure that the
1378UTF-8 flag is stripped off. Note that at the time of this writing
1379(October 2002) the mentioned modules are not UTF-8-aware. Please
1380check the documentation to verify if this is still true.
1381
1382  if ($] > 5.007) {
1383    require Encode;
1384    $val = Encode::encode_utf8($val); # make octets
1385  }
1386
1387=item *
1388
1389A scalar we got back from an extension
1390
1391If you believe the scalar comes back as UTF-8, you will most likely
1392want the UTF-8 flag restored:
1393
1394  if ($] > 5.007) {
1395    require Encode;
1396    $val = Encode::decode_utf8($val);
1397  }
1398
1399=item *
1400
1401Same thing, if you are really sure it is UTF-8
1402
1403  if ($] > 5.007) {
1404    require Encode;
1405    Encode::_utf8_on($val);
1406  }
1407
1408=item *
1409
1410A wrapper for fetchrow_array and fetchrow_hashref
1411
1412When the database contains only UTF-8, a wrapper function or method is
1413a convenient way to replace all your fetchrow_array and
1414fetchrow_hashref calls. A wrapper function will also make it easier to
1415adapt to future enhancements in your database driver. Note that at the
1416time of this writing (October 2002), the DBI has no standardized way
1417to deal with UTF-8 data. Please check the documentation to verify if
1418that is still true.
1419
1420  sub fetchrow {
1421    my($self, $sth, $what) = @_; # $what is one of fetchrow_{array,hashref}
1422    if ($] < 5.007) {
1423      return $sth->$what;
1424    } else {
1425      require Encode;
1426      if (wantarray) {
1427        my @arr = $sth->$what;
1428        for (@arr) {
1429          defined && /[^\000-\177]/ && Encode::_utf8_on($_);
1430        }
1431        return @arr;
1432      } else {
1433        my $ret = $sth->$what;
1434        if (ref $ret) {
1435          for my $k (keys %$ret) {
1436            defined && /[^\000-\177]/ && Encode::_utf8_on($_) for $ret->{$k};
1437          }
1438          return $ret;
1439        } else {
1440          defined && /[^\000-\177]/ && Encode::_utf8_on($_) for $ret;
1441          return $ret;
1442        }
1443      }
1444    }
1445  }
1446
1447
1448=item *
1449
1450A large scalar that you know can only contain ASCII
1451
1452Scalars that contain only ASCII and are marked as UTF-8 are sometimes
1453a drag to your program. If you recognize such a situation, just remove
1454the UTF-8 flag:
1455
1456  utf8::downgrade($val) if $] > 5.007;
1457
1458=back
1459
1460=head1 SEE ALSO
1461
1462L<perluniintro>, L<encoding>, L<Encode>, L<open>, L<utf8>, L<bytes>,
1463L<perlretut>, L<perlvar/"${^UNICODE}">
1464
1465=cut
1466