This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
[patch] :utf8 updates
[perl5.git] / pod / perlunicode.pod
CommitLineData
393fec97
GS
1=head1 NAME
2
3perlunicode - Unicode support in Perl
4
5=head1 DESCRIPTION
6
0a1f2d14 7=head2 Important Caveats
21bad921 8
376d9008 9Unicode support is an extensive requirement. While Perl does not
c349b1b9
JH
10implement the Unicode standard or the accompanying technical reports
11from cover to cover, Perl does support many Unicode features.
21bad921 12
2575c402
JW
13People who want to learn to use Unicode in Perl, should probably read
14L<the Perl Unicode tutorial|perlunitut> before reading this reference
15document.
16
13a2d996 17=over 4
21bad921 18
fae2c0fb 19=item Input and Output Layers
21bad921 20
376d9008 21Perl knows when a filehandle uses Perl's internal Unicode encodings
1bfb14c4
JH
22(UTF-8, or UTF-EBCDIC if in EBCDIC) if the filehandle is opened with
23the ":utf8" layer. Other encodings can be converted to Perl's
24encoding on input or from Perl's encoding on output by use of the
25":encoding(...)" layer. See L<open>.
c349b1b9 26
2575c402 27To indicate that Perl source itself is in UTF-8, use C<use utf8;>.
21bad921
GS
28
29=item Regular Expressions
30
c349b1b9 31The regular expression compiler produces polymorphic opcodes. That is,
376d9008 32the pattern adapts to the data and automatically switches to the Unicode
2575c402
JW
33character scheme when presented with data that is internally encoded in
34UTF-8 -- or instead uses a traditional byte scheme when presented with
35byte data.
21bad921 36
ad0029c4 37=item C<use utf8> still needed to enable UTF-8/UTF-EBCDIC in scripts
21bad921 38
376d9008
JB
39As a compatibility measure, the C<use utf8> pragma must be explicitly
40included to enable recognition of UTF-8 in the Perl scripts themselves
1bfb14c4
JH
41(in string or regular expression literals, or in identifier names) on
42ASCII-based machines or to recognize UTF-EBCDIC on EBCDIC-based
376d9008 43machines. B<These are the only times when an explicit C<use utf8>
8f8cf39c 44is needed.> See L<utf8>.
21bad921 45
7aa207d6
JH
46=item BOM-marked scripts and UTF-16 scripts autodetected
47
48If a Perl script begins marked with the Unicode BOM (UTF-16LE, UTF16-BE,
49or UTF-8), or if the script looks like non-BOM-marked UTF-16 of either
50endianness, Perl will correctly read in the script as Unicode.
51(BOMless UTF-8 cannot be effectively recognized or differentiated from
52ISO 8859-1 or other eight-bit encodings.)
53
990e18f7
AT
54=item C<use encoding> needed to upgrade non-Latin-1 byte strings
55
38a44b82 56By default, there is a fundamental asymmetry in Perl's Unicode model:
990e18f7
AT
57implicit upgrading from byte strings to Unicode strings assumes that
58they were encoded in I<ISO 8859-1 (Latin-1)>, but Unicode strings are
59downgraded with UTF-8 encoding. This happens because the first 256
60codepoints in Unicode happens to agree with Latin-1.
61
990e18f7
AT
62See L</"Byte and Character Semantics"> for more details.
63
21bad921
GS
64=back
65
376d9008 66=head2 Byte and Character Semantics
393fec97 67
376d9008 68Beginning with version 5.6, Perl uses logically-wide characters to
3e4dbfed 69represent strings internally.
393fec97 70
376d9008
JB
71In future, Perl-level operations will be expected to work with
72characters rather than bytes.
393fec97 73
376d9008 74However, as an interim compatibility measure, Perl aims to
75daf61c
JH
75provide a safe migration path from byte semantics to character
76semantics for programs. For operations where Perl can unambiguously
376d9008 77decide that the input data are characters, Perl switches to
75daf61c
JH
78character semantics. For operations where this determination cannot
79be made without additional information from the user, Perl decides in
376d9008 80favor of compatibility and chooses to use byte semantics.
8cbd9a7a
GS
81
82This behavior preserves compatibility with earlier versions of Perl,
376d9008
JB
83which allowed byte semantics in Perl operations only if
84none of the program's inputs were marked as being as source of Unicode
8cbd9a7a
GS
85character data. Such data may come from filehandles, from calls to
86external programs, from information provided by the system (such as %ENV),
21bad921 87or from literals and constants in the source text.
8cbd9a7a 88
376d9008
JB
89The C<bytes> pragma will always, regardless of platform, force byte
90semantics in a particular lexical scope. See L<bytes>.
8cbd9a7a
GS
91
92The C<utf8> pragma is primarily a compatibility device that enables
75daf61c 93recognition of UTF-(8|EBCDIC) in literals encountered by the parser.
376d9008
JB
94Note that this pragma is only required while Perl defaults to byte
95semantics; when character semantics become the default, this pragma
96may become a no-op. See L<utf8>.
97
98Unless explicitly stated, Perl operators use character semantics
99for Unicode data and byte semantics for non-Unicode data.
100The decision to use character semantics is made transparently. If
101input data comes from a Unicode source--for example, if a character
fae2c0fb 102encoding layer is added to a filehandle or a literal Unicode
376d9008
JB
103string constant appears in a program--character semantics apply.
104Otherwise, byte semantics are in effect. The C<bytes> pragma should
105be used to force byte semantics on Unicode data.
106
107If strings operating under byte semantics and strings with Unicode
990e18f7
AT
108character data are concatenated, the new string will be created by
109decoding the byte strings as I<ISO 8859-1 (Latin-1)>, even if the
110old Unicode string used EBCDIC. This translation is done without
2575c402 111regard to the system's native 8-bit encoding.
7dedd01f 112
feda178f 113Under character semantics, many operations that formerly operated on
376d9008 114bytes now operate on characters. A character in Perl is
feda178f 115logically just a number ranging from 0 to 2**31 or so. Larger
376d9008
JB
116characters may encode into longer sequences of bytes internally, but
117this internal detail is mostly hidden for Perl code.
118See L<perluniintro> for more.
393fec97 119
376d9008 120=head2 Effects of Character Semantics
393fec97
GS
121
122Character semantics have the following effects:
123
124=over 4
125
126=item *
127
376d9008 128Strings--including hash keys--and regular expression patterns may
574c8022 129contain characters that have an ordinal value larger than 255.
393fec97 130
2575c402
JW
131If you use a Unicode editor to edit your program, Unicode characters may
132occur directly within the literal strings in UTF-8 encoding, or UTF-16.
133(The former requires a BOM or C<use utf8>, the latter requires a BOM.)
3e4dbfed 134
2575c402
JW
135Unicode characters can also be added to a string by using the C<\x{...}>
136notation. The Unicode code for the desired character, in hexadecimal,
137should be placed in the braces. For instance, a smiley face is
138C<\x{263A}>. This encoding scheme only works for all characters, but
139for characters under 0x100, note that Perl may use an 8 bit encoding
140internally, for optimization and/or backward compatibility.
3e4dbfed
JF
141
142Additionally, if you
574c8022 143
3e4dbfed 144 use charnames ':full';
574c8022 145
1bfb14c4
JH
146you can use the C<\N{...}> notation and put the official Unicode
147character name within the braces, such as C<\N{WHITE SMILING FACE}>.
376d9008 148
393fec97
GS
149=item *
150
574c8022
JH
151If an appropriate L<encoding> is specified, identifiers within the
152Perl script may contain Unicode alphanumeric characters, including
376d9008
JB
153ideographs. Perl does not currently attempt to canonicalize variable
154names.
393fec97 155
393fec97
GS
156=item *
157
1bfb14c4 158Regular expressions match characters instead of bytes. "." matches
2575c402 159a character instead of a byte.
393fec97 160
393fec97
GS
161=item *
162
163Character classes in regular expressions match characters instead of
376d9008 164bytes and match against the character properties specified in the
1bfb14c4 165Unicode properties database. C<\w> can be used to match a Japanese
75daf61c 166ideograph, for instance.
393fec97 167
393fec97
GS
168=item *
169
eb0cc9e3 170Named Unicode properties, scripts, and block ranges may be used like
376d9008 171character classes via the C<\p{}> "matches property" construct and
822502e5
ST
172the C<\P{}> negation, "doesn't match property".
173
2575c402 174See L</"Unicode Character Properties"> for more details.
822502e5
ST
175
176You can define your own character properties and use them
177in the regular expression with the C<\p{}> or C<\P{}> construct.
178
179See L</"User-Defined Character Properties"> for more details.
180
181=item *
182
183The special pattern C<\X> matches any extended Unicode
184sequence--"a combining character sequence" in Standardese--where the
185first character is a base character and subsequent characters are mark
186characters that apply to the base character. C<\X> is equivalent to
187C<(?:\PM\pM*)>.
188
189=item *
190
191The C<tr///> operator translates characters instead of bytes. Note
192that the C<tr///CU> functionality has been removed. For similar
193functionality see pack('U0', ...) and pack('C0', ...).
194
195=item *
196
197Case translation operators use the Unicode case translation tables
198when character input is provided. Note that C<uc()>, or C<\U> in
199interpolated strings, translates to uppercase, while C<ucfirst>,
200or C<\u> in interpolated strings, translates to titlecase in languages
201that make the distinction.
202
203=item *
204
205Most operators that deal with positions or lengths in a string will
206automatically switch to using character positions, including
207C<chop()>, C<chomp()>, C<substr()>, C<pos()>, C<index()>, C<rindex()>,
208C<sprintf()>, C<write()>, and C<length()>. An operator that
209specifically does not switch is C<vec()>. Operators that really don't
210care include operators that treat strings as a bucket of bits such as
211C<sort()>, and operators dealing with filenames.
212
213=item *
214
215The C<pack()>/C<unpack()> letter C<C> does I<not> change, since it is often
216used for byte-oriented formats. Again, think C<char> in the C language.
217
218There is a new C<U> specifier that converts between Unicode characters
219and code points. There is also a C<W> specifier that is the equivalent of
220C<chr>/C<ord> and properly handles character values even if they are above 255.
221
222=item *
223
224The C<chr()> and C<ord()> functions work on characters, similar to
225C<pack("W")> and C<unpack("W")>, I<not> C<pack("C")> and
226C<unpack("C")>. C<pack("C")> and C<unpack("C")> are methods for
227emulating byte-oriented C<chr()> and C<ord()> on Unicode strings.
228While these methods reveal the internal encoding of Unicode strings,
229that is not something one normally needs to care about at all.
230
231=item *
232
233The bit string operators, C<& | ^ ~>, can operate on character data.
234However, for backward compatibility, such as when using bit string
235operations when characters are all less than 256 in ordinal value, one
236should not use C<~> (the bit complement) with characters of both
237values less than 256 and values greater than 256. Most importantly,
238DeMorgan's laws (C<~($x|$y) eq ~$x&~$y> and C<~($x&$y) eq ~$x|~$y>)
239will not hold. The reason for this mathematical I<faux pas> is that
240the complement cannot return B<both> the 8-bit (byte-wide) bit
241complement B<and> the full character-wide bit complement.
242
243=item *
244
245lc(), uc(), lcfirst(), and ucfirst() work for the following cases:
246
247=over 8
248
249=item *
250
251the case mapping is from a single Unicode character to another
252single Unicode character, or
253
254=item *
255
256the case mapping is from a single Unicode character to more
257than one Unicode character.
258
259=back
260
261Things to do with locales (Lithuanian, Turkish, Azeri) do B<not> work
262since Perl does not understand the concept of Unicode locales.
263
264See the Unicode Technical Report #21, Case Mappings, for more details.
265
266But you can also define your own mappings to be used in the lc(),
267lcfirst(), uc(), and ucfirst() (or their string-inlined versions).
268
269See L</"User-Defined Case Mappings"> for more details.
270
271=back
272
273=over 4
274
275=item *
276
277And finally, C<scalar reverse()> reverses by character rather than by byte.
278
279=back
280
281=head2 Unicode Character Properties
282
283Named Unicode properties, scripts, and block ranges may be used like
284character classes via the C<\p{}> "matches property" construct and
285the C<\P{}> negation, "doesn't match property".
1bfb14c4
JH
286
287For instance, C<\p{Lu}> matches any character with the Unicode "Lu"
288(Letter, uppercase) property, while C<\p{M}> matches any character
289with an "M" (mark--accents and such) property. Brackets are not
290required for single letter properties, so C<\p{M}> is equivalent to
291C<\pM>. Many predefined properties are available, such as
292C<\p{Mirrored}> and C<\p{Tibetan}>.
4193bef7 293
cfc01aea 294The official Unicode script and block names have spaces and dashes as
376d9008 295separators, but for convenience you can use dashes, spaces, or
1bfb14c4
JH
296underbars, and case is unimportant. It is recommended, however, that
297for consistency you use the following naming: the official Unicode
298script, property, or block name (see below for the additional rules
299that apply to block names) with whitespace and dashes removed, and the
300words "uppercase-first-lowercase-rest". C<Latin-1 Supplement> thus
301becomes C<Latin1Supplement>.
4193bef7 302
376d9008
JB
303You can also use negation in both C<\p{}> and C<\P{}> by introducing a caret
304(^) between the first brace and the property name: C<\p{^Tamil}> is
eb0cc9e3 305equal to C<\P{Tamil}>.
4193bef7 306
14bb0a9a 307B<NOTE: the properties, scripts, and blocks listed here are as of
8158862b 308Unicode 5.0.0 in July 2006.>
14bb0a9a 309
822502e5
ST
310=over 4
311
312=item General Category
313
eb0cc9e3 314Here are the basic Unicode General Category properties, followed by their
68cd2d32 315long form. You can use either; C<\p{Lu}> and C<\p{UppercaseLetter}>,
376d9008 316for instance, are identical.
393fec97 317
d73e5302
JH
318 Short Long
319
320 L Letter
12ac2576 321 LC CasedLetter
eb0cc9e3
JH
322 Lu UppercaseLetter
323 Ll LowercaseLetter
324 Lt TitlecaseLetter
325 Lm ModifierLetter
326 Lo OtherLetter
d73e5302
JH
327
328 M Mark
eb0cc9e3
JH
329 Mn NonspacingMark
330 Mc SpacingMark
331 Me EnclosingMark
d73e5302
JH
332
333 N Number
eb0cc9e3
JH
334 Nd DecimalNumber
335 Nl LetterNumber
336 No OtherNumber
d73e5302
JH
337
338 P Punctuation
eb0cc9e3
JH
339 Pc ConnectorPunctuation
340 Pd DashPunctuation
341 Ps OpenPunctuation
342 Pe ClosePunctuation
343 Pi InitialPunctuation
d73e5302 344 (may behave like Ps or Pe depending on usage)
eb0cc9e3 345 Pf FinalPunctuation
d73e5302 346 (may behave like Ps or Pe depending on usage)
eb0cc9e3 347 Po OtherPunctuation
d73e5302
JH
348
349 S Symbol
eb0cc9e3
JH
350 Sm MathSymbol
351 Sc CurrencySymbol
352 Sk ModifierSymbol
353 So OtherSymbol
d73e5302
JH
354
355 Z Separator
eb0cc9e3
JH
356 Zs SpaceSeparator
357 Zl LineSeparator
358 Zp ParagraphSeparator
d73e5302
JH
359
360 C Other
e150c829
JH
361 Cc Control
362 Cf Format
eb0cc9e3
JH
363 Cs Surrogate (not usable)
364 Co PrivateUse
e150c829 365 Cn Unassigned
1ac13f9a 366
376d9008 367Single-letter properties match all characters in any of the
3e4dbfed 368two-letter sub-properties starting with the same letter.
12ac2576
JP
369C<LC> and C<L&> are special cases, which are aliases for the set of
370C<Ll>, C<Lu>, and C<Lt>.
32293815 371
eb0cc9e3 372Because Perl hides the need for the user to understand the internal
1bfb14c4
JH
373representation of Unicode characters, there is no need to implement
374the somewhat messy concept of surrogates. C<Cs> is therefore not
eb0cc9e3 375supported.
d73e5302 376
822502e5
ST
377=item Bidirectional Character Types
378
376d9008 379Because scripts differ in their directionality--Hebrew is
12ac2576
JP
380written right to left, for example--Unicode supplies these properties in
381the BidiClass class:
32293815 382
eb0cc9e3 383 Property Meaning
92e830a9 384
12ac2576
JP
385 L Left-to-Right
386 LRE Left-to-Right Embedding
387 LRO Left-to-Right Override
388 R Right-to-Left
389 AL Right-to-Left Arabic
390 RLE Right-to-Left Embedding
391 RLO Right-to-Left Override
392 PDF Pop Directional Format
393 EN European Number
394 ES European Number Separator
395 ET European Number Terminator
396 AN Arabic Number
397 CS Common Number Separator
398 NSM Non-Spacing Mark
399 BN Boundary Neutral
400 B Paragraph Separator
401 S Segment Separator
402 WS Whitespace
403 ON Other Neutrals
404
405For example, C<\p{BidiClass:R}> matches characters that are normally
eb0cc9e3
JH
406written right to left.
407
822502e5 408=item Scripts
2796c109 409
376d9008
JB
410The script names which can be used by C<\p{...}> and C<\P{...}>,
411such as in C<\p{Latin}> or C<\p{Cyrillic}>, are as follows:
2796c109 412
1ac13f9a 413 Arabic
e9ad1727 414 Armenian
8158862b 415 Balinese
1ac13f9a 416 Bengali
e9ad1727 417 Bopomofo
8158862b
ST
418 Braille
419 Buginese
1d81abf3 420 Buhid
eb0cc9e3 421 CanadianAboriginal
e9ad1727 422 Cherokee
8158862b
ST
423 Coptic
424 Cuneiform
425 Cypriot
e9ad1727
JH
426 Cyrillic
427 Deseret
428 Devanagari
429 Ethiopic
430 Georgian
8158862b 431 Glagolitic
e9ad1727
JH
432 Gothic
433 Greek
1ac13f9a 434 Gujarati
e9ad1727
JH
435 Gurmukhi
436 Han
437 Hangul
1d81abf3 438 Hanunoo
e9ad1727
JH
439 Hebrew
440 Hiragana
441 Inherited
1ac13f9a 442 Kannada
e9ad1727 443 Katakana
8158862b 444 Kharoshthi
e9ad1727 445 Khmer
1ac13f9a 446 Lao
e9ad1727 447 Latin
8158862b
ST
448 Limbu
449 LinearB
e9ad1727
JH
450 Malayalam
451 Mongolian
1ac13f9a 452 Myanmar
8158862b
ST
453 NewTaiLue
454 Nko
1ac13f9a 455 Ogham
eb0cc9e3 456 OldItalic
8158862b 457 OldPersian
e9ad1727 458 Oriya
8158862b
ST
459 Osmanya
460 PhagsPa
461 Phoenician
1ac13f9a 462 Runic
8158862b 463 Shavian
e9ad1727 464 Sinhala
8158862b 465 SylotiNagri
e9ad1727 466 Syriac
1d81abf3
JH
467 Tagalog
468 Tagbanwa
8158862b 469 TaiLe
e9ad1727
JH
470 Tamil
471 Telugu
472 Thaana
473 Thai
474 Tibetan
8158862b
ST
475 Tifinagh
476 Ugaritic
1ac13f9a 477 Yi
1ac13f9a 478
822502e5
ST
479=item Extended property classes
480
376d9008 481Extended property classes can supplement the basic
1ac13f9a
JH
482properties, defined by the F<PropList> Unicode database:
483
1d81abf3 484 ASCIIHexDigit
eb0cc9e3 485 BidiControl
1ac13f9a 486 Dash
1d81abf3 487 Deprecated
1ac13f9a
JH
488 Diacritic
489 Extender
eb0cc9e3 490 HexDigit
e9ad1727
JH
491 Hyphen
492 Ideographic
1d81abf3
JH
493 IDSBinaryOperator
494 IDSTrinaryOperator
eb0cc9e3 495 JoinControl
1d81abf3 496 LogicalOrderException
eb0cc9e3
JH
497 NoncharacterCodePoint
498 OtherAlphabetic
1d81abf3
JH
499 OtherDefaultIgnorableCodePoint
500 OtherGraphemeExtend
8158862b
ST
501 OtherIDStart
502 OtherIDContinue
eb0cc9e3
JH
503 OtherLowercase
504 OtherMath
505 OtherUppercase
8158862b
ST
506 PatternSyntax
507 PatternWhiteSpace
eb0cc9e3 508 QuotationMark
1d81abf3
JH
509 Radical
510 SoftDotted
8158862b 511 STerm
1d81abf3
JH
512 TerminalPunctuation
513 UnifiedIdeograph
8158862b 514 VariationSelector
eb0cc9e3 515 WhiteSpace
1ac13f9a 516
376d9008 517and there are further derived properties:
1ac13f9a 518
8158862b
ST
519 Alphabetic = Lu + Ll + Lt + Lm + Lo + Nl + OtherAlphabetic
520 Lowercase = Ll + OtherLowercase
521 Uppercase = Lu + OtherUppercase
522 Math = Sm + OtherMath
1ac13f9a 523
8158862b
ST
524 IDStart = Lu + Ll + Lt + Lm + Lo + Nl + OtherIDStart
525 IDContinue = IDStart + Mn + Mc + Nd + Pc + OtherIDContinue
1ac13f9a 526
8158862b
ST
527 DefaultIgnorableCodePoint
528 = OtherDefaultIgnorableCodePoint
529 + Cf + Cc + Cs + Noncharacters + VariationSelector
530 - WhiteSpace - FFF9..FFFB (Annotation Characters)
531
532 Any = Any code points (i.e. U+0000 to U+10FFFF)
533 Assigned = Any non-Cn code points (i.e. synonym for \P{Cn})
534 Unassigned = Synonym for \p{Cn}
535 ASCII = ASCII (i.e. U+0000 to U+007F)
536
537 Common = Any character (or unassigned code point)
538 not explicitly assigned to a script
2796c109 539
822502e5
ST
540=item Use of "Is" Prefix
541
1bfb14c4
JH
542For backward compatibility (with Perl 5.6), all properties mentioned
543so far may have C<Is> prepended to their name, so C<\P{IsLu}>, for
544example, is equal to C<\P{Lu}>.
eb0cc9e3 545
822502e5 546=item Blocks
2796c109 547
1bfb14c4
JH
548In addition to B<scripts>, Unicode also defines B<blocks> of
549characters. The difference between scripts and blocks is that the
550concept of scripts is closer to natural languages, while the concept
551of blocks is more of an artificial grouping based on groups of 256
376d9008 552Unicode characters. For example, the C<Latin> script contains letters
1bfb14c4 553from many blocks but does not contain all the characters from those
376d9008
JB
554blocks. It does not, for example, contain digits, because digits are
555shared across many scripts. Digits and similar groups, like
556punctuation, are in a category called C<Common>.
2796c109 557
8158862b 558For more about scripts, see the UAX#24 "Script Names":
cfc01aea 559
8158862b 560 http://www.unicode.org/reports/tr24/
cfc01aea
JF
561
562For more about blocks, see:
563
564 http://www.unicode.org/Public/UNIDATA/Blocks.txt
2796c109 565
376d9008
JB
566Block names are given with the C<In> prefix. For example, the
567Katakana block is referenced via C<\p{InKatakana}>. The C<In>
7eabb34d 568prefix may be omitted if there is no naming conflict with a script
eb0cc9e3 569or any other property, but it is recommended that C<In> always be used
1bfb14c4 570for block tests to avoid confusion.
eb0cc9e3
JH
571
572These block names are supported:
573
8158862b 574 InAegeanNumbers
1d81abf3 575 InAlphabeticPresentationForms
8158862b
ST
576 InAncientGreekMusicalNotation
577 InAncientGreekNumbers
1d81abf3
JH
578 InArabic
579 InArabicPresentationFormsA
580 InArabicPresentationFormsB
8158862b 581 InArabicSupplement
1d81abf3
JH
582 InArmenian
583 InArrows
8158862b 584 InBalinese
1d81abf3
JH
585 InBasicLatin
586 InBengali
587 InBlockElements
588 InBopomofo
589 InBopomofoExtended
590 InBoxDrawing
591 InBraillePatterns
8158862b 592 InBuginese
1d81abf3
JH
593 InBuhid
594 InByzantineMusicalSymbols
595 InCJKCompatibility
596 InCJKCompatibilityForms
597 InCJKCompatibilityIdeographs
598 InCJKCompatibilityIdeographsSupplement
599 InCJKRadicalsSupplement
8158862b 600 InCJKStrokes
1d81abf3
JH
601 InCJKSymbolsAndPunctuation
602 InCJKUnifiedIdeographs
603 InCJKUnifiedIdeographsExtensionA
604 InCJKUnifiedIdeographsExtensionB
605 InCherokee
606 InCombiningDiacriticalMarks
8158862b 607 InCombiningDiacriticalMarksSupplement
1d81abf3
JH
608 InCombiningDiacriticalMarksforSymbols
609 InCombiningHalfMarks
610 InControlPictures
8158862b
ST
611 InCoptic
612 InCountingRodNumerals
613 InCuneiform
614 InCuneiformNumbersAndPunctuation
1d81abf3 615 InCurrencySymbols
8158862b 616 InCypriotSyllabary
1d81abf3 617 InCyrillic
8158862b 618 InCyrillicSupplement
1d81abf3
JH
619 InDeseret
620 InDevanagari
621 InDingbats
622 InEnclosedAlphanumerics
623 InEnclosedCJKLettersAndMonths
624 InEthiopic
8158862b
ST
625 InEthiopicExtended
626 InEthiopicSupplement
1d81abf3
JH
627 InGeneralPunctuation
628 InGeometricShapes
629 InGeorgian
8158862b
ST
630 InGeorgianSupplement
631 InGlagolitic
1d81abf3
JH
632 InGothic
633 InGreekExtended
634 InGreekAndCoptic
635 InGujarati
636 InGurmukhi
637 InHalfwidthAndFullwidthForms
638 InHangulCompatibilityJamo
639 InHangulJamo
640 InHangulSyllables
641 InHanunoo
642 InHebrew
643 InHighPrivateUseSurrogates
644 InHighSurrogates
645 InHiragana
646 InIPAExtensions
647 InIdeographicDescriptionCharacters
648 InKanbun
649 InKangxiRadicals
650 InKannada
651 InKatakana
652 InKatakanaPhoneticExtensions
8158862b 653 InKharoshthi
1d81abf3 654 InKhmer
8158862b 655 InKhmerSymbols
1d81abf3
JH
656 InLao
657 InLatin1Supplement
658 InLatinExtendedA
659 InLatinExtendedAdditional
660 InLatinExtendedB
8158862b
ST
661 InLatinExtendedC
662 InLatinExtendedD
1d81abf3 663 InLetterlikeSymbols
8158862b
ST
664 InLimbu
665 InLinearBIdeograms
666 InLinearBSyllabary
1d81abf3
JH
667 InLowSurrogates
668 InMalayalam
669 InMathematicalAlphanumericSymbols
670 InMathematicalOperators
671 InMiscellaneousMathematicalSymbolsA
672 InMiscellaneousMathematicalSymbolsB
673 InMiscellaneousSymbols
8158862b 674 InMiscellaneousSymbolsAndArrows
1d81abf3 675 InMiscellaneousTechnical
8158862b 676 InModifierToneLetters
1d81abf3
JH
677 InMongolian
678 InMusicalSymbols
679 InMyanmar
8158862b
ST
680 InNKo
681 InNewTaiLue
1d81abf3
JH
682 InNumberForms
683 InOgham
684 InOldItalic
8158862b 685 InOldPersian
1d81abf3
JH
686 InOpticalCharacterRecognition
687 InOriya
8158862b
ST
688 InOsmanya
689 InPhagspa
690 InPhoenician
691 InPhoneticExtensions
692 InPhoneticExtensionsSupplement
1d81abf3
JH
693 InPrivateUseArea
694 InRunic
8158862b 695 InShavian
1d81abf3
JH
696 InSinhala
697 InSmallFormVariants
698 InSpacingModifierLetters
699 InSpecials
700 InSuperscriptsAndSubscripts
701 InSupplementalArrowsA
702 InSupplementalArrowsB
703 InSupplementalMathematicalOperators
8158862b 704 InSupplementalPunctuation
1d81abf3
JH
705 InSupplementaryPrivateUseAreaA
706 InSupplementaryPrivateUseAreaB
8158862b 707 InSylotiNagri
1d81abf3
JH
708 InSyriac
709 InTagalog
710 InTagbanwa
711 InTags
8158862b
ST
712 InTaiLe
713 InTaiXuanJingSymbols
1d81abf3
JH
714 InTamil
715 InTelugu
716 InThaana
717 InThai
718 InTibetan
8158862b
ST
719 InTifinagh
720 InUgaritic
1d81abf3
JH
721 InUnifiedCanadianAboriginalSyllabics
722 InVariationSelectors
8158862b
ST
723 InVariationSelectorsSupplement
724 InVerticalForms
1d81abf3
JH
725 InYiRadicals
726 InYiSyllables
8158862b 727 InYijingHexagramSymbols
32293815 728
393fec97
GS
729=back
730
376d9008 731=head2 User-Defined Character Properties
491fd90a
JH
732
733You can define your own character properties by defining subroutines
bac0b425
JP
734whose names begin with "In" or "Is". The subroutines can be defined in
735any package. The user-defined properties can be used in the regular
736expression C<\p> and C<\P> constructs; if you are using a user-defined
737property from a package other than the one you are in, you must specify
738its package in the C<\p> or C<\P> construct.
739
740 # assuming property IsForeign defined in Lang::
741 package main; # property package name required
742 if ($txt =~ /\p{Lang::IsForeign}+/) { ... }
743
744 package Lang; # property package name not required
745 if ($txt =~ /\p{IsForeign}+/) { ... }
746
747
748Note that the effect is compile-time and immutable once defined.
491fd90a 749
376d9008
JB
750The subroutines must return a specially-formatted string, with one
751or more newline-separated lines. Each line must be one of the following:
491fd90a
JH
752
753=over 4
754
755=item *
756
510254c9
A
757A single hexadecimal number denoting a Unicode code point to include.
758
759=item *
760
99a6b1f0 761Two hexadecimal numbers separated by horizontal whitespace (space or
376d9008 762tabular characters) denoting a range of Unicode code points to include.
491fd90a
JH
763
764=item *
765
376d9008 766Something to include, prefixed by "+": a built-in character
bac0b425
JP
767property (prefixed by "utf8::") or a user-defined character property,
768to represent all the characters in that property; two hexadecimal code
769points for a range; or a single hexadecimal code point.
491fd90a
JH
770
771=item *
772
376d9008 773Something to exclude, prefixed by "-": an existing character
bac0b425
JP
774property (prefixed by "utf8::") or a user-defined character property,
775to represent all the characters in that property; two hexadecimal code
776points for a range; or a single hexadecimal code point.
491fd90a
JH
777
778=item *
779
376d9008 780Something to negate, prefixed "!": an existing character
bac0b425
JP
781property (prefixed by "utf8::") or a user-defined character property,
782to represent all the characters in that property; two hexadecimal code
783points for a range; or a single hexadecimal code point.
784
785=item *
786
787Something to intersect with, prefixed by "&": an existing character
788property (prefixed by "utf8::") or a user-defined character property,
789for all the characters except the characters in the property; two
790hexadecimal code points for a range; or a single hexadecimal code point.
491fd90a
JH
791
792=back
793
794For example, to define a property that covers both the Japanese
795syllabaries (hiragana and katakana), you can define
796
797 sub InKana {
d5822f25
A
798 return <<END;
799 3040\t309F
800 30A0\t30FF
491fd90a
JH
801 END
802 }
803
d5822f25
A
804Imagine that the here-doc end marker is at the beginning of the line.
805Now you can use C<\p{InKana}> and C<\P{InKana}>.
491fd90a
JH
806
807You could also have used the existing block property names:
808
809 sub InKana {
810 return <<'END';
811 +utf8::InHiragana
812 +utf8::InKatakana
813 END
814 }
815
816Suppose you wanted to match only the allocated characters,
d5822f25 817not the raw block ranges: in other words, you want to remove
491fd90a
JH
818the non-characters:
819
820 sub InKana {
821 return <<'END';
822 +utf8::InHiragana
823 +utf8::InKatakana
824 -utf8::IsCn
825 END
826 }
827
828The negation is useful for defining (surprise!) negated classes.
829
830 sub InNotKana {
831 return <<'END';
832 !utf8::InHiragana
833 -utf8::InKatakana
834 +utf8::IsCn
835 END
836 }
837
bac0b425
JP
838Intersection is useful for getting the common characters matched by
839two (or more) classes.
840
841 sub InFooAndBar {
842 return <<'END';
843 +main::Foo
844 &main::Bar
845 END
846 }
847
848It's important to remember not to use "&" for the first set -- that
849would be intersecting with nothing (resulting in an empty set).
850
822502e5
ST
851=head2 User-Defined Case Mappings
852
3a2263fe
RGS
853You can also define your own mappings to be used in the lc(),
854lcfirst(), uc(), and ucfirst() (or their string-inlined versions).
822502e5
ST
855The principle is similar to that of user-defined character
856properties: to define subroutines in the C<main> package
3a2263fe
RGS
857with names like C<ToLower> (for lc() and lcfirst()), C<ToTitle> (for
858the first character in ucfirst()), and C<ToUpper> (for uc(), and the
859rest of the characters in ucfirst()).
860
861The string returned by the subroutines needs now to be three
862hexadecimal numbers separated by tabulators: start of the source
863range, end of the source range, and start of the destination range.
864For example:
865
866 sub ToUpper {
867 return <<END;
868 0061\t0063\t0041
869 END
870 }
871
872defines an uc() mapping that causes only the characters "a", "b", and
873"c" to be mapped to "A", "B", "C", all other characters will remain
874unchanged.
875
876If there is no source range to speak of, that is, the mapping is from
877a single character to another single character, leave the end of the
878source range empty, but the two tabulator characters are still needed.
879For example:
880
881 sub ToLower {
882 return <<END;
883 0041\t\t0061
884 END
885 }
886
887defines a lc() mapping that causes only "A" to be mapped to "a", all
888other characters will remain unchanged.
889
890(For serious hackers only) If you want to introspect the default
891mappings, you can find the data in the directory
892C<$Config{privlib}>/F<unicore/To/>. The mapping data is returned as
893the here-document, and the C<utf8::ToSpecFoo> are special exception
894mappings derived from <$Config{privlib}>/F<unicore/SpecialCasing.txt>.
895The C<Digit> and C<Fold> mappings that one can see in the directory
896are not directly user-accessible, one can use either the
897C<Unicode::UCD> module, or just match case-insensitively (that's when
898the C<Fold> mapping is used).
899
822502e5
ST
900A final note on the user-defined case mappings: they will be used
901only if the scalar has been marked as having Unicode characters.
902Old byte-style strings will not be affected.
3a2263fe 903
376d9008 904=head2 Character Encodings for Input and Output
8cbd9a7a 905
7221edc9 906See L<Encode>.
8cbd9a7a 907
c29a771d 908=head2 Unicode Regular Expression Support Level
776f8809 909
376d9008
JB
910The following list of Unicode support for regular expressions describes
911all the features currently supported. The references to "Level N"
8158862b
ST
912and the section numbers refer to the Unicode Technical Standard #18,
913"Unicode Regular Expressions", version 11, in May 2005.
776f8809
JH
914
915=over 4
916
917=item *
918
919Level 1 - Basic Unicode Support
920
8158862b
ST
921 RL1.1 Hex Notation - done [1]
922 RL1.2 Properties - done [2][3]
923 RL1.2a Compatibility Properties - done [4]
924 RL1.3 Subtraction and Intersection - MISSING [5]
925 RL1.4 Simple Word Boundaries - done [6]
926 RL1.5 Simple Loose Matches - done [7]
927 RL1.6 Line Boundaries - MISSING [8]
928 RL1.7 Supplementary Code Points - done [9]
929
930 [1] \x{...}
931 [2] \p{...} \P{...}
932 [3] supports not only minimal list (general category, scripts,
933 Alphabetic, Lowercase, Uppercase, WhiteSpace,
934 NoncharacterCodePoint, DefaultIgnorableCodePoint, Any,
935 ASCII, Assigned), but also bidirectional types, blocks, etc.
936 (see L</"Unicode Character Properties">)
937 [4] \d \D \s \S \w \W \X [:prop:] [:^prop:]
938 [5] can use regular expression look-ahead [a] or
939 user-defined character properties [b] to emulate set operations
940 [6] \b \B
941 [7] note that Perl does Full case-folding in matching, not Simple:
835863de 942 for example U+1F88 is equivalent with U+1F00 U+03B9,
e0f9d4a8 943 not with 1F80. This difference matters for certain Greek
376d9008
JB
944 capital letters with certain modifiers: the Full case-folding
945 decomposes the letter, while the Simple case-folding would map
e0f9d4a8 946 it to a single character.
8158862b
ST
947 [8] should do ^ and $ also on U+000B (\v in C), FF (\f), CR (\r),
948 CRLF (\r\n), NEL (U+0085), LS (U+2028), and PS (U+2029);
949 should also affect <>, $., and script line numbers;
950 should not split lines within CRLF [c] (i.e. there is no empty
951 line between \r and \n)
952 [9] UTF-8/UTF-EBDDIC used in perl allows not only U+10000 to U+10FFFF
953 but also beyond U+10FFFF [d]
7207e29d 954
237bad5b 955[a] You can mimic class subtraction using lookahead.
8158862b 956For example, what UTS#18 might write as
29bdacb8 957
dbe420b4
JH
958 [{Greek}-[{UNASSIGNED}]]
959
960in Perl can be written as:
961
1d81abf3
JH
962 (?!\p{Unassigned})\p{InGreekAndCoptic}
963 (?=\p{Assigned})\p{InGreekAndCoptic}
dbe420b4
JH
964
965But in this particular example, you probably really want
966
1bfb14c4 967 \p{GreekAndCoptic}
dbe420b4
JH
968
969which will match assigned characters known to be part of the Greek script.
29bdacb8 970
5ca1ac52 971Also see the Unicode::Regex::Set module, it does implement the full
8158862b
ST
972UTS#18 grouping, intersection, union, and removal (subtraction) syntax.
973
974[b] '+' for union, '-' for removal (set-difference), '&' for intersection
975(see L</"User-Defined Character Properties">)
976
977[c] Try the C<:crlf> layer (see L<PerlIO>).
5ca1ac52 978
8158862b
ST
979[d] Avoid C<use warning 'utf8';> (or say C<no warning 'utf8';>) to allow
980U+FFFF (C<\x{FFFF}>).
237bad5b 981
776f8809
JH
982=item *
983
984Level 2 - Extended Unicode Support
985
8158862b
ST
986 RL2.1 Canonical Equivalents - MISSING [10][11]
987 RL2.2 Default Grapheme Clusters - MISSING [12][13]
988 RL2.3 Default Word Boundaries - MISSING [14]
989 RL2.4 Default Loose Matches - MISSING [15]
990 RL2.5 Name Properties - MISSING [16]
991 RL2.6 Wildcard Properties - MISSING
992
993 [10] see UAX#15 "Unicode Normalization Forms"
994 [11] have Unicode::Normalize but not integrated to regexes
995 [12] have \X but at this level . should equal that
996 [13] UAX#29 "Text Boundaries" considers CRLF and Hangul syllable
997 clusters as a single grapheme cluster.
998 [14] see UAX#29, Word Boundaries
999 [15] see UAX#21 "Case Mappings"
1000 [16] have \N{...} but neither compute names of CJK Ideographs
1001 and Hangul Syllables nor use a loose match [e]
1002
1003[e] C<\N{...}> allows namespaces (see L<charnames>).
776f8809
JH
1004
1005=item *
1006
8158862b
ST
1007Level 3 - Tailored Support
1008
1009 RL3.1 Tailored Punctuation - MISSING
1010 RL3.2 Tailored Grapheme Clusters - MISSING [17][18]
1011 RL3.3 Tailored Word Boundaries - MISSING
1012 RL3.4 Tailored Loose Matches - MISSING
1013 RL3.5 Tailored Ranges - MISSING
1014 RL3.6 Context Matching - MISSING [19]
1015 RL3.7 Incremental Matches - MISSING
1016 ( RL3.8 Unicode Set Sharing )
1017 RL3.9 Possible Match Sets - MISSING
1018 RL3.10 Folded Matching - MISSING [20]
1019 RL3.11 Submatchers - MISSING
1020
1021 [17] see UAX#10 "Unicode Collation Algorithms"
1022 [18] have Unicode::Collate but not integrated to regexes
1023 [19] have (?<=x) and (?=x), but look-aheads or look-behinds should see
1024 outside of the target substring
1025 [20] need insensitive matching for linguistic features other than case;
1026 for example, hiragana to katakana, wide and narrow, simplified Han
1027 to traditional Han (see UTR#30 "Character Foldings")
776f8809
JH
1028
1029=back
1030
c349b1b9
JH
1031=head2 Unicode Encodings
1032
376d9008
JB
1033Unicode characters are assigned to I<code points>, which are abstract
1034numbers. To use these numbers, various encodings are needed.
c349b1b9
JH
1035
1036=over 4
1037
c29a771d 1038=item *
5cb3728c
RB
1039
1040UTF-8
c349b1b9 1041
3e4dbfed 1042UTF-8 is a variable-length (1 to 6 bytes, current character allocations
376d9008
JB
1043require 4 bytes), byte-order independent encoding. For ASCII (and we
1044really do mean 7-bit ASCII, not another 8-bit encoding), UTF-8 is
1045transparent.
c349b1b9 1046
8c007b5a 1047The following table is from Unicode 3.2.
05632f9a
JH
1048
1049 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
1050
8c007b5a
JH
1051 U+0000..U+007F 00..7F
1052 U+0080..U+07FF C2..DF 80..BF
ec90690f
ST
1053 U+0800..U+0FFF E0 A0..BF 80..BF
1054 U+1000..U+CFFF E1..EC 80..BF 80..BF
1055 U+D000..U+D7FF ED 80..9F 80..BF
8c007b5a 1056 U+D800..U+DFFF ******* ill-formed *******
ec90690f 1057 U+E000..U+FFFF EE..EF 80..BF 80..BF
05632f9a
JH
1058 U+10000..U+3FFFF F0 90..BF 80..BF 80..BF
1059 U+40000..U+FFFFF F1..F3 80..BF 80..BF 80..BF
1060 U+100000..U+10FFFF F4 80..8F 80..BF 80..BF
1061
376d9008
JB
1062Note the C<A0..BF> in C<U+0800..U+0FFF>, the C<80..9F> in
1063C<U+D000...U+D7FF>, the C<90..B>F in C<U+10000..U+3FFFF>, and the
1064C<80...8F> in C<U+100000..U+10FFFF>. The "gaps" are caused by legal
1065UTF-8 avoiding non-shortest encodings: it is technically possible to
1066UTF-8-encode a single code point in different ways, but that is
1067explicitly forbidden, and the shortest possible encoding should always
1068be used. So that's what Perl does.
37361303 1069
376d9008 1070Another way to look at it is via bits:
05632f9a
JH
1071
1072 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
1073
1074 0aaaaaaa 0aaaaaaa
1075 00000bbbbbaaaaaa 110bbbbb 10aaaaaa
1076 ccccbbbbbbaaaaaa 1110cccc 10bbbbbb 10aaaaaa
1077 00000dddccccccbbbbbbaaaaaa 11110ddd 10cccccc 10bbbbbb 10aaaaaa
1078
1079As you can see, the continuation bytes all begin with C<10>, and the
8c007b5a 1080leading bits of the start byte tell how many bytes the are in the
05632f9a
JH
1081encoded character.
1082
c29a771d 1083=item *
5cb3728c
RB
1084
1085UTF-EBCDIC
dbe420b4 1086
376d9008 1087Like UTF-8 but EBCDIC-safe, in the way that UTF-8 is ASCII-safe.
dbe420b4 1088
c29a771d 1089=item *
5cb3728c 1090
1e54db1a 1091UTF-16, UTF-16BE, UTF-16LE, Surrogates, and BOMs (Byte Order Marks)
c349b1b9 1092
1bfb14c4
JH
1093The followings items are mostly for reference and general Unicode
1094knowledge, Perl doesn't use these constructs internally.
dbe420b4 1095
c349b1b9 1096UTF-16 is a 2 or 4 byte encoding. The Unicode code points
1bfb14c4
JH
1097C<U+0000..U+FFFF> are stored in a single 16-bit unit, and the code
1098points C<U+10000..U+10FFFF> in two 16-bit units. The latter case is
c349b1b9
JH
1099using I<surrogates>, the first 16-bit unit being the I<high
1100surrogate>, and the second being the I<low surrogate>.
1101
376d9008 1102Surrogates are code points set aside to encode the C<U+10000..U+10FFFF>
c349b1b9 1103range of Unicode code points in pairs of 16-bit units. The I<high
376d9008
JB
1104surrogates> are the range C<U+D800..U+DBFF>, and the I<low surrogates>
1105are the range C<U+DC00..U+DFFF>. The surrogate encoding is
c349b1b9
JH
1106
1107 $hi = ($uni - 0x10000) / 0x400 + 0xD800;
1108 $lo = ($uni - 0x10000) % 0x400 + 0xDC00;
1109
1110and the decoding is
1111
1a3fa709 1112 $uni = 0x10000 + ($hi - 0xD800) * 0x400 + ($lo - 0xDC00);
c349b1b9 1113
feda178f 1114If you try to generate surrogates (for example by using chr()), you
376d9008
JB
1115will get a warning if warnings are turned on, because those code
1116points are not valid for a Unicode character.
9466bab6 1117
376d9008 1118Because of the 16-bitness, UTF-16 is byte-order dependent. UTF-16
c349b1b9 1119itself can be used for in-memory computations, but if storage or
376d9008
JB
1120transfer is required either UTF-16BE (big-endian) or UTF-16LE
1121(little-endian) encodings must be chosen.
c349b1b9
JH
1122
1123This introduces another problem: what if you just know that your data
376d9008
JB
1124is UTF-16, but you don't know which endianness? Byte Order Marks, or
1125BOMs, are a solution to this. A special character has been reserved
86bbd6d1 1126in Unicode to function as a byte order marker: the character with the
376d9008 1127code point C<U+FEFF> is the BOM.
042da322 1128
c349b1b9 1129The trick is that if you read a BOM, you will know the byte order,
376d9008
JB
1130since if it was written on a big-endian platform, you will read the
1131bytes C<0xFE 0xFF>, but if it was written on a little-endian platform,
1132you will read the bytes C<0xFF 0xFE>. (And if the originating platform
1133was writing in UTF-8, you will read the bytes C<0xEF 0xBB 0xBF>.)
042da322 1134
86bbd6d1 1135The way this trick works is that the character with the code point
376d9008
JB
1136C<U+FFFE> is guaranteed not to be a valid Unicode character, so the
1137sequence of bytes C<0xFF 0xFE> is unambiguously "BOM, represented in
1bfb14c4 1138little-endian format" and cannot be C<U+FFFE>, represented in big-endian
042da322 1139format".
c349b1b9 1140
c29a771d 1141=item *
5cb3728c 1142
1e54db1a 1143UTF-32, UTF-32BE, UTF-32LE
c349b1b9
JH
1144
1145The UTF-32 family is pretty much like the UTF-16 family, expect that
042da322 1146the units are 32-bit, and therefore the surrogate scheme is not
376d9008
JB
1147needed. The BOM signatures will be C<0x00 0x00 0xFE 0xFF> for BE and
1148C<0xFF 0xFE 0x00 0x00> for LE.
c349b1b9 1149
c29a771d 1150=item *
5cb3728c
RB
1151
1152UCS-2, UCS-4
c349b1b9 1153
86bbd6d1 1154Encodings defined by the ISO 10646 standard. UCS-2 is a 16-bit
376d9008 1155encoding. Unlike UTF-16, UCS-2 is not extensible beyond C<U+FFFF>,
339cfa0e
JH
1156because it does not use surrogates. UCS-4 is a 32-bit encoding,
1157functionally identical to UTF-32.
c349b1b9 1158
c29a771d 1159=item *
5cb3728c
RB
1160
1161UTF-7
c349b1b9 1162
376d9008
JB
1163A seven-bit safe (non-eight-bit) encoding, which is useful if the
1164transport or storage is not eight-bit safe. Defined by RFC 2152.
c349b1b9 1165
95a1a48b
JH
1166=back
1167
0d7c09bb
JH
1168=head2 Security Implications of Unicode
1169
1170=over 4
1171
1172=item *
1173
1174Malformed UTF-8
bf0fa0b2
JH
1175
1176Unfortunately, the specification of UTF-8 leaves some room for
1177interpretation of how many bytes of encoded output one should generate
376d9008
JB
1178from one input Unicode character. Strictly speaking, the shortest
1179possible sequence of UTF-8 bytes should be generated,
1180because otherwise there is potential for an input buffer overflow at
feda178f 1181the receiving end of a UTF-8 connection. Perl always generates the
376d9008
JB
1182shortest length UTF-8, and with warnings on Perl will warn about
1183non-shortest length UTF-8 along with other malformations, such as the
1184surrogates, which are not real Unicode code points.
bf0fa0b2 1185
0d7c09bb
JH
1186=item *
1187
1188Regular expressions behave slightly differently between byte data and
376d9008
JB
1189character (Unicode) data. For example, the "word character" character
1190class C<\w> will work differently depending on if data is eight-bit bytes
1191or Unicode.
0d7c09bb 1192
376d9008
JB
1193In the first case, the set of C<\w> characters is either small--the
1194default set of alphabetic characters, digits, and the "_"--or, if you
0d7c09bb
JH
1195are using a locale (see L<perllocale>), the C<\w> might contain a few
1196more letters according to your language and country.
1197
376d9008 1198In the second case, the C<\w> set of characters is much, much larger.
1bfb14c4
JH
1199Most importantly, even in the set of the first 256 characters, it will
1200probably match different characters: unlike most locales, which are
1201specific to a language and country pair, Unicode classifies all the
1202characters that are letters I<somewhere> as C<\w>. For example, your
1203locale might not think that LATIN SMALL LETTER ETH is a letter (unless
1204you happen to speak Icelandic), but Unicode does.
0d7c09bb 1205
376d9008 1206As discussed elsewhere, Perl has one foot (two hooves?) planted in
1bfb14c4
JH
1207each of two worlds: the old world of bytes and the new world of
1208characters, upgrading from bytes to characters when necessary.
376d9008
JB
1209If your legacy code does not explicitly use Unicode, no automatic
1210switch-over to characters should happen. Characters shouldn't get
1bfb14c4
JH
1211downgraded to bytes, either. It is possible to accidentally mix bytes
1212and characters, however (see L<perluniintro>), in which case C<\w> in
1213regular expressions might start behaving differently. Review your
1214code. Use warnings and the C<strict> pragma.
0d7c09bb
JH
1215
1216=back
1217
c349b1b9
JH
1218=head2 Unicode in Perl on EBCDIC
1219
376d9008
JB
1220The way Unicode is handled on EBCDIC platforms is still
1221experimental. On such platforms, references to UTF-8 encoding in this
1222document and elsewhere should be read as meaning the UTF-EBCDIC
1223specified in Unicode Technical Report 16, unless ASCII vs. EBCDIC issues
c349b1b9 1224are specifically discussed. There is no C<utfebcdic> pragma or
376d9008 1225":utfebcdic" layer; rather, "utf8" and ":utf8" are reused to mean
86bbd6d1
PN
1226the platform's "natural" 8-bit encoding of Unicode. See L<perlebcdic>
1227for more discussion of the issues.
c349b1b9 1228
b310b053
JH
1229=head2 Locales
1230
4616122b 1231Usually locale settings and Unicode do not affect each other, but
b310b053
JH
1232there are a couple of exceptions:
1233
1234=over 4
1235
1236=item *
1237
8aa8f774
JH
1238You can enable automatic UTF-8-ification of your standard file
1239handles, default C<open()> layer, and C<@ARGV> by using either
1240the C<-C> command line switch or the C<PERL_UNICODE> environment
1241variable, see L<perlrun> for the documentation of the C<-C> switch.
b310b053
JH
1242
1243=item *
1244
376d9008
JB
1245Perl tries really hard to work both with Unicode and the old
1246byte-oriented world. Most often this is nice, but sometimes Perl's
1247straddling of the proverbial fence causes problems.
b310b053
JH
1248
1249=back
1250
1aad1664
JH
1251=head2 When Unicode Does Not Happen
1252
1253While Perl does have extensive ways to input and output in Unicode,
1254and few other 'entry points' like the @ARGV which can be interpreted
1255as Unicode (UTF-8), there still are many places where Unicode (in some
1256encoding or another) could be given as arguments or received as
1257results, or both, but it is not.
1258
6cd4dd6c
JH
1259The following are such interfaces. For all of these interfaces Perl
1260currently (as of 5.8.3) simply assumes byte strings both as arguments
1261and results, or UTF-8 strings if the C<encoding> pragma has been used.
1aad1664
JH
1262
1263One reason why Perl does not attempt to resolve the role of Unicode in
1264this cases is that the answers are highly dependent on the operating
1265system and the file system(s). For example, whether filenames can be
1266in Unicode, and in exactly what kind of encoding, is not exactly a
1267portable concept. Similarly for the qx and system: how well will the
1268'command line interface' (and which of them?) handle Unicode?
1269
1270=over 4
1271
557a2462
RB
1272=item *
1273
254c2b64 1274chdir, chmod, chown, chroot, exec, link, lstat, mkdir,
1e8e8236 1275rename, rmdir, stat, symlink, truncate, unlink, utime, -X
557a2462
RB
1276
1277=item *
1278
1279%ENV
1280
1281=item *
1282
1283glob (aka the <*>)
1284
1285=item *
1aad1664 1286
557a2462 1287open, opendir, sysopen
1aad1664 1288
557a2462 1289=item *
1aad1664 1290
557a2462 1291qx (aka the backtick operator), system
1aad1664 1292
557a2462 1293=item *
1aad1664 1294
557a2462 1295readdir, readlink
1aad1664
JH
1296
1297=back
1298
1299=head2 Forcing Unicode in Perl (Or Unforcing Unicode in Perl)
1300
1301Sometimes (see L</"When Unicode Does Not Happen">) there are
1302situations where you simply need to force Perl to believe that a byte
1303string is UTF-8, or vice versa. The low-level calls
1304utf8::upgrade($bytestring) and utf8::downgrade($utf8string) are
1305the answers.
1306
1307Do not use them without careful thought, though: Perl may easily get
1308very confused, angry, or even crash, if you suddenly change the 'nature'
1309of scalar like that. Especially careful you have to be if you use the
1310utf8::upgrade(): any random byte string is not valid UTF-8.
1311
95a1a48b
JH
1312=head2 Using Unicode in XS
1313
3a2263fe
RGS
1314If you want to handle Perl Unicode in XS extensions, you may find the
1315following C APIs useful. See also L<perlguts/"Unicode Support"> for an
1316explanation about Unicode at the XS level, and L<perlapi> for the API
1317details.
95a1a48b
JH
1318
1319=over 4
1320
1321=item *
1322
1bfb14c4
JH
1323C<DO_UTF8(sv)> returns true if the C<UTF8> flag is on and the bytes
1324pragma is not in effect. C<SvUTF8(sv)> returns true is the C<UTF8>
1325flag is on; the bytes pragma is ignored. The C<UTF8> flag being on
1326does B<not> mean that there are any characters of code points greater
1327than 255 (or 127) in the scalar or that there are even any characters
1328in the scalar. What the C<UTF8> flag means is that the sequence of
1329octets in the representation of the scalar is the sequence of UTF-8
1330encoded code points of the characters of a string. The C<UTF8> flag
1331being off means that each octet in this representation encodes a
1332single character with code point 0..255 within the string. Perl's
1333Unicode model is not to use UTF-8 until it is absolutely necessary.
95a1a48b
JH
1334
1335=item *
1336
fb9cc174 1337C<uvuni_to_utf8(buf, chr)> writes a Unicode character code point into
1bfb14c4 1338a buffer encoding the code point as UTF-8, and returns a pointer
95a1a48b
JH
1339pointing after the UTF-8 bytes.
1340
1341=item *
1342
376d9008
JB
1343C<utf8_to_uvuni(buf, lenp)> reads UTF-8 encoded bytes from a buffer and
1344returns the Unicode character code point and, optionally, the length of
1345the UTF-8 byte sequence.
95a1a48b
JH
1346
1347=item *
1348
376d9008
JB
1349C<utf8_length(start, end)> returns the length of the UTF-8 encoded buffer
1350in characters. C<sv_len_utf8(sv)> returns the length of the UTF-8 encoded
95a1a48b
JH
1351scalar.
1352
1353=item *
1354
376d9008
JB
1355C<sv_utf8_upgrade(sv)> converts the string of the scalar to its UTF-8
1356encoded form. C<sv_utf8_downgrade(sv)> does the opposite, if
1357possible. C<sv_utf8_encode(sv)> is like sv_utf8_upgrade except that
1358it does not set the C<UTF8> flag. C<sv_utf8_decode()> does the
1359opposite of C<sv_utf8_encode()>. Note that none of these are to be
1360used as general-purpose encoding or decoding interfaces: C<use Encode>
1361for that. C<sv_utf8_upgrade()> is affected by the encoding pragma
1362but C<sv_utf8_downgrade()> is not (since the encoding pragma is
1363designed to be a one-way street).
95a1a48b
JH
1364
1365=item *
1366
376d9008 1367C<is_utf8_char(s)> returns true if the pointer points to a valid UTF-8
90f968e0 1368character.
95a1a48b
JH
1369
1370=item *
1371
376d9008 1372C<is_utf8_string(buf, len)> returns true if C<len> bytes of the buffer
95a1a48b
JH
1373are valid UTF-8.
1374
1375=item *
1376
376d9008
JB
1377C<UTF8SKIP(buf)> will return the number of bytes in the UTF-8 encoded
1378character in the buffer. C<UNISKIP(chr)> will return the number of bytes
1379required to UTF-8-encode the Unicode character code point. C<UTF8SKIP()>
90f968e0 1380is useful for example for iterating over the characters of a UTF-8
376d9008 1381encoded buffer; C<UNISKIP()> is useful, for example, in computing
90f968e0 1382the size required for a UTF-8 encoded buffer.
95a1a48b
JH
1383
1384=item *
1385
376d9008 1386C<utf8_distance(a, b)> will tell the distance in characters between the
95a1a48b
JH
1387two pointers pointing to the same UTF-8 encoded buffer.
1388
1389=item *
1390
376d9008
JB
1391C<utf8_hop(s, off)> will return a pointer to an UTF-8 encoded buffer
1392that is C<off> (positive or negative) Unicode characters displaced
1393from the UTF-8 buffer C<s>. Be careful not to overstep the buffer:
1394C<utf8_hop()> will merrily run off the end or the beginning of the
1395buffer if told to do so.
95a1a48b 1396
d2cc3551
JH
1397=item *
1398
376d9008
JB
1399C<pv_uni_display(dsv, spv, len, pvlim, flags)> and
1400C<sv_uni_display(dsv, ssv, pvlim, flags)> are useful for debugging the
1401output of Unicode strings and scalars. By default they are useful
1402only for debugging--they display B<all> characters as hexadecimal code
1bfb14c4
JH
1403points--but with the flags C<UNI_DISPLAY_ISPRINT>,
1404C<UNI_DISPLAY_BACKSLASH>, and C<UNI_DISPLAY_QQ> you can make the
1405output more readable.
d2cc3551
JH
1406
1407=item *
1408
376d9008
JB
1409C<ibcmp_utf8(s1, pe1, u1, l1, u1, s2, pe2, l2, u2)> can be used to
1410compare two strings case-insensitively in Unicode. For case-sensitive
1411comparisons you can just use C<memEQ()> and C<memNE()> as usual.
d2cc3551 1412
c349b1b9
JH
1413=back
1414
95a1a48b
JH
1415For more information, see L<perlapi>, and F<utf8.c> and F<utf8.h>
1416in the Perl source code distribution.
1417
c29a771d
JH
1418=head1 BUGS
1419
376d9008 1420=head2 Interaction with Locales
7eabb34d 1421
376d9008
JB
1422Use of locales with Unicode data may lead to odd results. Currently,
1423Perl attempts to attach 8-bit locale info to characters in the range
14240..255, but this technique is demonstrably incorrect for locales that
1425use characters above that range when mapped into Unicode. Perl's
1426Unicode support will also tend to run slower. Use of locales with
1427Unicode is discouraged.
c29a771d 1428
376d9008 1429=head2 Interaction with Extensions
7eabb34d 1430
376d9008 1431When Perl exchanges data with an extension, the extension should be
2575c402 1432able to understand the UTF8 flag and act accordingly. If the
376d9008
JB
1433extension doesn't know about the flag, it's likely that the extension
1434will return incorrectly-flagged data.
7eabb34d
A
1435
1436So if you're working with Unicode data, consult the documentation of
1437every module you're using if there are any issues with Unicode data
1438exchange. If the documentation does not talk about Unicode at all,
a73d23f6 1439suspect the worst and probably look at the source to learn how the
376d9008 1440module is implemented. Modules written completely in Perl shouldn't
a73d23f6
RGS
1441cause problems. Modules that directly or indirectly access code written
1442in other programming languages are at risk.
7eabb34d 1443
376d9008 1444For affected functions, the simple strategy to avoid data corruption is
7eabb34d 1445to always make the encoding of the exchanged data explicit. Choose an
376d9008 1446encoding that you know the extension can handle. Convert arguments passed
7eabb34d
A
1447to the extensions to that encoding and convert results back from that
1448encoding. Write wrapper functions that do the conversions for you, so
1449you can later change the functions when the extension catches up.
1450
376d9008 1451To provide an example, let's say the popular Foo::Bar::escape_html
7eabb34d
A
1452function doesn't deal with Unicode data yet. The wrapper function
1453would convert the argument to raw UTF-8 and convert the result back to
376d9008 1454Perl's internal representation like so:
7eabb34d
A
1455
1456 sub my_escape_html ($) {
1457 my($what) = shift;
1458 return unless defined $what;
1459 Encode::decode_utf8(Foo::Bar::escape_html(Encode::encode_utf8($what)));
1460 }
1461
1462Sometimes, when the extension does not convert data but just stores
1463and retrieves them, you will be in a position to use the otherwise
1464dangerous Encode::_utf8_on() function. Let's say the popular
66b79f27 1465C<Foo::Bar> extension, written in C, provides a C<param> method that
7eabb34d
A
1466lets you store and retrieve data according to these prototypes:
1467
1468 $self->param($name, $value); # set a scalar
1469 $value = $self->param($name); # retrieve a scalar
1470
1471If it does not yet provide support for any encoding, one could write a
1472derived class with such a C<param> method:
1473
1474 sub param {
1475 my($self,$name,$value) = @_;
1476 utf8::upgrade($name); # make sure it is UTF-8 encoded
af55fc6a 1477 if (defined $value) {
7eabb34d
A
1478 utf8::upgrade($value); # make sure it is UTF-8 encoded
1479 return $self->SUPER::param($name,$value);
1480 } else {
1481 my $ret = $self->SUPER::param($name);
1482 Encode::_utf8_on($ret); # we know, it is UTF-8 encoded
1483 return $ret;
1484 }
1485 }
1486
a73d23f6
RGS
1487Some extensions provide filters on data entry/exit points, such as
1488DB_File::filter_store_key and family. Look out for such filters in
66b79f27 1489the documentation of your extensions, they can make the transition to
7eabb34d
A
1490Unicode data much easier.
1491
376d9008 1492=head2 Speed
7eabb34d 1493
c29a771d 1494Some functions are slower when working on UTF-8 encoded strings than
574c8022 1495on byte encoded strings. All functions that need to hop over
7c17141f
JH
1496characters such as length(), substr() or index(), or matching regular
1497expressions can work B<much> faster when the underlying data are
1498byte-encoded.
1499
1500In Perl 5.8.0 the slowness was often quite spectacular; in Perl 5.8.1
1501a caching scheme was introduced which will hopefully make the slowness
a104b433
JH
1502somewhat less spectacular, at least for some operations. In general,
1503operations with UTF-8 encoded strings are still slower. As an example,
1504the Unicode properties (character classes) like C<\p{Nd}> are known to
1505be quite a bit slower (5-20 times) than their simpler counterparts
1506like C<\d> (then again, there 268 Unicode characters matching C<Nd>
1507compared with the 10 ASCII characters matching C<d>).
666f95b9 1508
c8d992ba
A
1509=head2 Porting code from perl-5.6.X
1510
1511Perl 5.8 has a different Unicode model from 5.6. In 5.6 the programmer
1512was required to use the C<utf8> pragma to declare that a given scope
1513expected to deal with Unicode data and had to make sure that only
1514Unicode data were reaching that scope. If you have code that is
1515working with 5.6, you will need some of the following adjustments to
1516your code. The examples are written such that the code will continue
1517to work under 5.6, so you should be safe to try them out.
1518
1519=over 4
1520
1521=item *
1522
1523A filehandle that should read or write UTF-8
1524
1525 if ($] > 5.007) {
740d4bb2 1526 binmode $fh, ":encoding(utf8)";
c8d992ba
A
1527 }
1528
1529=item *
1530
1531A scalar that is going to be passed to some extension
1532
1533Be it Compress::Zlib, Apache::Request or any extension that has no
1534mention of Unicode in the manpage, you need to make sure that the
2575c402 1535UTF8 flag is stripped off. Note that at the time of this writing
c8d992ba
A
1536(October 2002) the mentioned modules are not UTF-8-aware. Please
1537check the documentation to verify if this is still true.
1538
1539 if ($] > 5.007) {
1540 require Encode;
1541 $val = Encode::encode_utf8($val); # make octets
1542 }
1543
1544=item *
1545
1546A scalar we got back from an extension
1547
1548If you believe the scalar comes back as UTF-8, you will most likely
2575c402 1549want the UTF8 flag restored:
c8d992ba
A
1550
1551 if ($] > 5.007) {
1552 require Encode;
1553 $val = Encode::decode_utf8($val);
1554 }
1555
1556=item *
1557
1558Same thing, if you are really sure it is UTF-8
1559
1560 if ($] > 5.007) {
1561 require Encode;
1562 Encode::_utf8_on($val);
1563 }
1564
1565=item *
1566
1567A wrapper for fetchrow_array and fetchrow_hashref
1568
1569When the database contains only UTF-8, a wrapper function or method is
1570a convenient way to replace all your fetchrow_array and
1571fetchrow_hashref calls. A wrapper function will also make it easier to
1572adapt to future enhancements in your database driver. Note that at the
1573time of this writing (October 2002), the DBI has no standardized way
1574to deal with UTF-8 data. Please check the documentation to verify if
1575that is still true.
1576
1577 sub fetchrow {
1578 my($self, $sth, $what) = @_; # $what is one of fetchrow_{array,hashref}
1579 if ($] < 5.007) {
1580 return $sth->$what;
1581 } else {
1582 require Encode;
1583 if (wantarray) {
1584 my @arr = $sth->$what;
1585 for (@arr) {
1586 defined && /[^\000-\177]/ && Encode::_utf8_on($_);
1587 }
1588 return @arr;
1589 } else {
1590 my $ret = $sth->$what;
1591 if (ref $ret) {
1592 for my $k (keys %$ret) {
1593 defined && /[^\000-\177]/ && Encode::_utf8_on($_) for $ret->{$k};
1594 }
1595 return $ret;
1596 } else {
1597 defined && /[^\000-\177]/ && Encode::_utf8_on($_) for $ret;
1598 return $ret;
1599 }
1600 }
1601 }
1602 }
1603
1604
1605=item *
1606
1607A large scalar that you know can only contain ASCII
1608
1609Scalars that contain only ASCII and are marked as UTF-8 are sometimes
1610a drag to your program. If you recognize such a situation, just remove
2575c402 1611the UTF8 flag:
c8d992ba
A
1612
1613 utf8::downgrade($val) if $] > 5.007;
1614
1615=back
1616
393fec97
GS
1617=head1 SEE ALSO
1618
2575c402 1619L<perlunitut>, L<perluniintro>, L<Encode>, L<open>, L<utf8>, L<bytes>,
a05d7ebb 1620L<perlretut>, L<perlvar/"${^UNICODE}">
393fec97
GS
1621
1622=cut