3 perlunicode - Unicode support in Perl
7 =head2 Important Caveats
9 Unicode support is an extensive requirement. While perl does not
10 implement the Unicode standard or the accompanying technical reports
11 from cover to cover, Perl does support many Unicode features.
15 =item Input and Output Disciplines
17 A filehandle can be marked as containing perl's internal Unicode
18 encoding (UTF-8 or UTF-EBCDIC) by opening it with the ":utf8" layer.
19 Other encodings can be converted to perl's encoding on input, or from
20 perl's encoding on output by use of the ":encoding(...)" layer.
23 To mark the Perl source itself as being in a particular encoding,
26 =item Regular Expressions
28 The regular expression compiler produces polymorphic opcodes. That is,
29 the pattern adapts to the data and automatically switch to the Unicode
30 character scheme when presented with Unicode data, or a traditional
31 byte scheme when presented with byte data.
33 =item C<use utf8> still needed to enable UTF-8/UTF-EBCDIC in scripts
35 The C<utf8> pragma implements the tables used for Unicode support.
36 However, these tables are automatically loaded on demand, so the
37 C<utf8> pragma should not normally be used.
39 As a compatibility measure, this pragma must be explicitly used to
40 enable recognition of UTF-8 in the Perl scripts themselves on ASCII
41 based machines or recognize UTF-EBCDIC on EBCDIC based machines.
42 B<NOTE: this should be the only place where an explicit C<use utf8>
45 You can also use the C<encoding> pragma to change the default encoding
46 of the data in your script; see L<encoding>.
50 =head2 Byte and Character semantics
52 Beginning with version 5.6, Perl uses logically wide characters to
53 represent strings internally. This internal representation of strings
54 uses either the UTF-8 or the UTF-EBCDIC encoding.
56 In future, Perl-level operations can be expected to work with
57 characters rather than bytes, in general.
59 However, as strictly an interim compatibility measure, Perl aims to
60 provide a safe migration path from byte semantics to character
61 semantics for programs. For operations where Perl can unambiguously
62 decide that the input data is characters, Perl now switches to
63 character semantics. For operations where this determination cannot
64 be made without additional information from the user, Perl decides in
65 favor of compatibility, and chooses to use byte semantics.
67 This behavior preserves compatibility with earlier versions of Perl,
68 which allowed byte semantics in Perl operations, but only as long as
69 none of the program's inputs are marked as being as source of Unicode
70 character data. Such data may come from filehandles, from calls to
71 external programs, from information provided by the system (such as %ENV),
72 or from literals and constants in the source text.
74 On Windows platforms, if the C<-C> command line switch is used, (or the
75 ${^WIDE_SYSTEM_CALLS} global flag is set to C<1>), all system calls
76 will use the corresponding wide character APIs. Note that this is
77 currently only implemented on Windows since other platforms lack an
78 API standard on this area.
80 Regardless of the above, the C<bytes> pragma can always be used to
81 force byte semantics in a particular lexical scope. See L<bytes>.
83 The C<utf8> pragma is primarily a compatibility device that enables
84 recognition of UTF-(8|EBCDIC) in literals encountered by the parser.
85 Note that this pragma is only required until a future version of Perl
86 in which character semantics will become the default. This pragma may
87 then become a no-op. See L<utf8>.
89 Unless mentioned otherwise, Perl operators will use character semantics
90 when they are dealing with Unicode data, and byte semantics otherwise.
91 Thus, character semantics for these operations apply transparently; if
92 the input data came from a Unicode source (for example, by adding a
93 character encoding discipline to the filehandle whence it came, or a
94 literal UTF-8 string constant in the program), character semantics
95 apply; otherwise, byte semantics are in effect. To force byte semantics
96 on Unicode data, the C<bytes> pragma should be used.
98 Notice that if you concatenate strings with byte semantics and strings
99 with Unicode character data, the bytes will by default be upgraded
100 I<as if they were ISO 8859-1 (Latin-1)> (or if in EBCDIC, after a
101 translation to ISO 8859-1). To change this, use the C<encoding>
102 pragma, see L<encoding>.
104 Under character semantics, many operations that formerly operated on
105 bytes change to operating on characters. For ASCII data this makes no
106 difference, because UTF-8 stores ASCII in single bytes, but for any
107 character greater than C<chr(127)>, the character B<may> be stored in
108 a sequence of two or more bytes, all of which have the high bit set.
110 For C1 controls or Latin 1 characters on an EBCDIC platform the
111 character may be stored in a UTF-EBCDIC multi byte sequence. But by
112 and large, the user need not worry about this, because Perl hides it
113 from the user. A character in Perl is logically just a number ranging
114 from 0 to 2**32 or so. Larger characters encode to longer sequences
115 of bytes internally, but again, this is just an internal detail which
116 is hidden at the Perl level.
118 =head2 Effects of character semantics
120 Character semantics have the following effects:
126 Strings and patterns may contain characters that have an ordinal value
129 Presuming you use a Unicode editor to edit your program, such
130 characters will typically occur directly within the literal strings as
131 UTF-8 (or UTF-EBCDIC on EBCDIC platforms) characters, but you can also
132 specify a particular character with an extension of the C<\x>
133 notation. UTF-X characters are specified by putting the hexadecimal
134 code within curlies after the C<\x>. For instance, a Unicode smiley
139 Identifiers within the Perl script may contain Unicode alphanumeric
140 characters, including ideographs. (You are currently on your own when
141 it comes to using the canonical forms of characters--Perl doesn't
142 (yet) attempt to canonicalize variable names for you.)
146 Regular expressions match characters instead of bytes. For instance,
147 "." matches a character instead of a byte. (However, the C<\C> pattern
148 is provided to force a match a single byte ("C<char>" in C, hence C<\C>).)
152 Character classes in regular expressions match characters instead of
153 bytes, and match against the character properties specified in the
154 Unicode properties database. So C<\w> can be used to match an
155 ideograph, for instance.
159 Named Unicode properties and block ranges make be used as character
160 classes via the new C<\p{}> (matches property) and C<\P{}> (doesn't
161 match property) constructs. For instance, C<\p{Lu}> matches any
162 character with the Unicode uppercase property, while C<\p{M}> matches
163 any mark character. Single letter properties may omit the brackets,
164 so that can be written C<\pM> also. Many predefined character classes
165 are available, such as C<\p{IsMirrored}> and C<\p{InTibetan}>.
167 The C<\p{Is...}> test for "general properties" such as "letter",
168 "digit", while the C<\p{In...}> test for Unicode scripts and blocks.
170 The official Unicode script and block names have spaces and dashes and
171 separators, but for convenience you can have dashes, spaces, and
172 underbars at every word division, and you need not care about correct
173 casing. It is recommended, however, that for consistency you use the
174 following naming: the official Unicode script, block, or property name
175 (see below for the additional rules that apply to block names),
176 with whitespace and dashes replaced with underbar, and the words
177 "uppercase-first-lowercase-rest". That is, "Latin-1 Supplement"
178 becomes "Latin_1_Supplement".
180 You can also negate both C<\p{}> and C<\P{}> by introducing a caret
181 (^) between the first curly and the property name: C<\p{^In_Tamil}> is
182 equal to C<\P{In_Tamil}>.
184 The C<In> and C<Is> can be left out: C<\p{Greek}> is equal to
185 C<\p{In_Greek}>, C<\P{Pd}> is equal to C<\P{Pd}>.
207 Pc Connector_Punctuation
211 Pi Initial_Punctuation
212 (may behave like Ps or Pe depending on usage)
214 (may behave like Ps or Pe depending on usage)
226 Zp Paragraph_Separator
235 There's also C<L&> which is an alias for C<Ll>, C<Lu>, and C<Lt>.
237 The following reserved ranges have C<In> tests:
239 CJK_Ideograph_Extension_A
242 Non_Private_Use_High_Surrogate
243 Private_Use_High_Surrogate
246 CJK_Ideograph_Extension_B
250 For example C<"\x{AC00}" =~ \p{HangulSyllable}> will test true.
251 (Handling of surrogates is not implemented yet, because Perl
252 uses UTF-8 and not UTF-16 internally to represent Unicode.)
254 Additionally, because scripts differ in their directionality
255 (for example Hebrew is written right to left), all characters
256 have their directionality defined:
259 BidiLRE Left-to-Right Embedding
260 BidiLRO Left-to-Right Override
262 BidiAL Right-to-Left Arabic
263 BidiRLE Right-to-Left Embedding
264 BidiRLO Right-to-Left Override
265 BidiPDF Pop Directional Format
266 BidiEN European Number
267 BidiES European Number Separator
268 BidiET European Number Terminator
270 BidiCS Common Number Separator
271 BidiNSM Non-Spacing Mark
272 BidiBN Boundary Neutral
273 BidiB Paragraph Separator
274 BidiS Segment Separator
276 BidiON Other Neutrals
282 The scripts available for C<\p{In...}> and C<\P{In...}>, for example
283 \p{InCyrillic>, are as follows, for example C<\p{InLatin}> or C<\P{InHan}>:
326 There are also extended property classes that supplement the basic
327 properties, defined by the F<PropList> Unicode database:
338 Noncharacter_Code_Point
346 and further derived properties:
348 Alphabetic Lu + Ll + Lt + Lm + Lo + Other_Alphabetic
349 Lowercase Ll + Other_Lowercase
350 Uppercase Lu + Other_Uppercase
353 ID_Start Lu + Ll + Lt + Lm + Lo + Nl
354 ID_Continue ID_Start + Mn + Mc + Nd + Pc
357 Assigned Any non-Cn character
358 Common Any character (or unassigned code point)
359 not explicitly assigned to a script
363 In addition to B<scripts>, Unicode also defines B<blocks> of
364 characters. The difference between scripts and blocks is that the
365 scripts concept is closer to natural languages, while the blocks
366 concept is more an artificial grouping based on groups of 256 Unicode
367 characters. For example, the C<Latin> script contains letters from
368 many blocks. On the other hand, the C<Latin> script does not contain
369 all the characters from those blocks, it does not for example contain
370 digits because digits are shared across many scripts. Digits and
371 other similar groups, like punctuation, are in a category called
374 For more about scripts see the UTR #24:
375 http://www.unicode.org/unicode/reports/tr24/
376 For more about blocks see
377 http://www.unicode.org/Public/UNIDATA/Blocks.txt
379 Because there are overlaps in naming (there are, for example, both
380 a script called C<Katakana> and a block called C<Katakana>, the block
381 version has C<Block> appended to its name, C<\p{InKatakanaBlock}>.
383 Notice that this definition was introduced in Perl 5.8.0: in Perl
384 5.6 only the blocks were used; in Perl 5.8.0 scripts became the
385 preferential Unicode character class definition; this meant that
386 the definitions of some character classes changed (the ones in the
387 below list that have the C<Block> appended).
389 Alphabetic Presentation Forms
391 Arabic Presentation Forms-A
392 Arabic Presentation Forms-B
402 Byzantine Musical Symbols
404 CJK Compatibility Forms
405 CJK Compatibility Ideographs
406 CJK Compatibility Ideographs Supplement
407 CJK Radicals Supplement
408 CJK Symbols and Punctuation
409 CJK Unified Ideographs
410 CJK Unified Ideographs Extension A
411 CJK Unified Ideographs Extension B
413 Combining Diacritical Marks
415 Combining Marks for Symbols
422 Enclosed Alphanumerics
423 Enclosed CJK Letters and Months
433 Halfwidth and Fullwidth Forms
434 Hangul Compatibility Jamo
438 High Private Use Surrogates
442 Ideographic Description Characters
450 Latin Extended Additional
456 Mathematical Alphanumeric Symbols
457 Mathematical Operators
458 Miscellaneous Symbols
459 Miscellaneous Technical
466 Optical Character Recognition
472 Spacing Modifier Letters
474 Superscripts and Subscripts
482 Unified Canadian Aboriginal Syllabics
490 The special pattern C<\X> match matches any extended Unicode sequence
491 (a "combining character sequence" in Standardese), where the first
492 character is a base character and subsequent characters are mark
493 characters that apply to the base character. It is equivalent to
498 The C<tr///> operator translates characters instead of bytes. Note
499 that the C<tr///CU> functionality has been removed, as the interface
500 was a mistake. For similar functionality see pack('U0', ...) and
505 Case translation operators use the Unicode case translation tables
506 when provided character input. Note that C<uc()> (also known as C<\U>
507 in doublequoted strings) translates to uppercase, while C<ucfirst>
508 (also known as C<\u> in doublequoted strings) translates to titlecase
509 (for languages that make the distinction). Naturally the
510 corresponding backslash sequences have the same semantics.
514 Most operators that deal with positions or lengths in the string will
515 automatically switch to using character positions, including
516 C<chop()>, C<substr()>, C<pos()>, C<index()>, C<rindex()>,
517 C<sprintf()>, C<write()>, and C<length()>. Operators that
518 specifically don't switch include C<vec()>, C<pack()>, and
519 C<unpack()>. Operators that really don't care include C<chomp()>, as
520 well as any other operator that treats a string as a bucket of bits,
521 such as C<sort()>, and the operators dealing with filenames.
525 The C<pack()>/C<unpack()> letters "C<c>" and "C<C>" do I<not> change,
526 since they're often used for byte-oriented formats. (Again, think
527 "C<char>" in the C language.) However, there is a new "C<U>" specifier
528 that will convert between UTF-8 characters and integers. (It works
529 outside of the utf8 pragma too.)
533 The C<chr()> and C<ord()> functions work on characters. This is like
534 C<pack("U")> and C<unpack("U")>, not like C<pack("C")> and
535 C<unpack("C")>. In fact, the latter are how you now emulate
536 byte-oriented C<chr()> and C<ord()> for Unicode strings.
537 (Note that this reveals the internal UTF-8 encoding of strings and
538 you are not supposed to do that unless you know what you are doing.)
542 The bit string operators C<& | ^ ~> can operate on character data.
543 However, for backward compatibility reasons (bit string operations
544 when the characters all are less than 256 in ordinal value) one should
545 not mix C<~> (the bit complement) and characters both less than 256 and
546 equal or greater than 256. Most importantly, the DeMorgan's laws
547 (C<~($x|$y) eq ~$x&~$y>, C<~($x&$y) eq ~$x|~$y>) won't hold.
548 Another way to look at this is that the complement cannot return
549 B<both> the 8-bit (byte) wide bit complement B<and> the full character
554 lc(), uc(), lcfirst(), and ucfirst() work for the following cases:
560 the case mapping is from a single Unicode character to another
561 single Unicode character
565 the case mapping is from a single Unicode character to more
566 than one Unicode character
570 What doesn't yet work are the following cases:
576 the "final sigma" (Greek)
580 anything to with locales (Lithuanian, Turkish, Azeri)
584 See the Unicode Technical Report #21, Case Mappings, for more details.
588 And finally, C<scalar reverse()> reverses by character rather than by byte.
592 =head2 Character encodings for input and output
598 As of yet, there is no method for automatically coercing input and
599 output to some encoding other than UTF-8 or UTF-EBCDIC. This is planned
600 in the near future, however.
602 Whether an arbitrary piece of data will be treated as "characters" or
603 "bytes" by internal operations cannot be divined at the current time.
605 Use of locales with utf8 may lead to odd results. Currently there is
606 some attempt to apply 8-bit locale info to characters in the range
607 0..255, but this is demonstrably incorrect for locales that use
608 characters above that range (when mapped into Unicode). It will also
609 tend to run slower. Avoidance of locales is strongly encouraged.
611 =head1 UNICODE REGULAR EXPRESSION SUPPORT LEVEL
613 The following list of Unicode regular expression support describes
614 feature by feature the Unicode support implemented in Perl as of Perl
615 5.8.0. The "Level N" and the section numbers refer to the Unicode
616 Technical Report 18, "Unicode Regular Expression Guidelines".
622 Level 1 - Basic Unicode Support
624 2.1 Hex Notation - done [1]
625 Named Notation - done [2]
626 2.2 Categories - done [3][4]
627 2.3 Subtraction - MISSING [5][6]
628 2.4 Simple Word Boundaries - done [7]
629 2.5 Simple Loose Matches - done [8]
630 2.6 End of Line - MISSING [9][10]
634 [ 3] . \p{Is...} \P{Is...}
635 [ 4] now scripts (see UTR#24 Script Names) in addition to blocks
637 [ 6] can use look-ahead to emulate subtraction (*)
638 [ 7] include Letters in word characters
639 [ 8] see UTR#21 Case Mappings: Perl implements 1:1 mappings
640 [ 9] see UTR#13 Unicode Newline Guidelines
641 [10] should do ^ and $ also on \x{2028} and \x{2029}
643 (*) Instead of [\u0370-\u03FF-[{UNASSIGNED}]] as suggested by the TR
644 18 you can use negated lookahead: to match currently assigned modern
645 Greek characters use for example
647 /(?!\p{Cn})[\x{0370}-\x{03ff}]/
649 In other words: the matched character must not be a non-assigned
650 character, but it must be in the block of modern Greek characters.
654 Level 2 - Extended Unicode Support
656 3.1 Surrogates - MISSING
657 3.2 Canonical Equivalents - MISSING [11][12]
658 3.3 Locale-Independent Graphemes - MISSING [13]
659 3.4 Locale-Independent Words - MISSING [14]
660 3.5 Locale-Independent Loose Matches - MISSING [15]
662 [11] see UTR#15 Unicode Normalization
663 [12] have Unicode::Normalize but not integrated to regexes
664 [13] have \X but at this level . should equal that
665 [14] need three classes, not just \w and \W
666 [15] see UTR#21 Case Mappings
670 Level 3 - Locale-Sensitive Support
672 4.1 Locale-Dependent Categories - MISSING
673 4.2 Locale-Dependent Graphemes - MISSING [16][17]
674 4.3 Locale-Dependent Words - MISSING
675 4.4 Locale-Dependent Loose Matches - MISSING
676 4.5 Locale-Dependent Ranges - MISSING
678 [16] see UTR#10 Unicode Collation Algorithms
679 [17] have Unicode::Collate but not integrated to regexes
683 =head2 Unicode Encodings
685 Unicode characters are assigned to I<code points> which are abstract
686 numbers. To use these numbers various encodings are needed.
692 UTF-8 is the encoding used internally by Perl. UTF-8 is a variable
693 length (1 to 6 bytes, current character allocations require 4 bytes),
694 byteorder independent encoding. For ASCII, UTF-8 is transparent
695 (and we really do mean 7-bit ASCII, not any 8-bit encoding).
697 The following table is from Unicode 3.1.
699 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
701 U+0000..U+007F 00..7F
702 U+0080..U+07FF C2..DF 80..BF
703 U+0800..U+0FFF E0 A0..BF 80..BF
704 U+1000..U+FFFF E1..EF 80..BF 80..BF
705 U+10000..U+3FFFF F0 90..BF 80..BF 80..BF
706 U+40000..U+FFFFF F1..F3 80..BF 80..BF 80..BF
707 U+100000..U+10FFFF F4 80..8F 80..BF 80..BF
709 Or, another way to look at it, as bits:
711 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
714 00000bbbbbaaaaaa 110bbbbb 10aaaaaa
715 ccccbbbbbbaaaaaa 1110cccc 10bbbbbb 10aaaaaa
716 00000dddccccccbbbbbbaaaaaa 11110ddd 10cccccc 10bbbbbb 10aaaaaa
718 As you can see, the continuation bytes all begin with C<10>, and the
719 leading bits of the start byte tells how many bytes the are in the
722 =item UTF-16, UTF-16BE, UTF16-LE, Surrogates, and BOMs (Byte Order Marks)
724 UTF-16 is a 2 or 4 byte encoding. The Unicode code points
725 0x0000..0xFFFF are stored in two 16-bit units, and the code points
726 0x010000..0x10FFFF in four 16-bit units. The latter case is
727 using I<surrogates>, the first 16-bit unit being the I<high
728 surrogate>, and the second being the I<low surrogate>.
730 Surrogates are code points set aside to encode the 0x01000..0x10FFFF
731 range of Unicode code points in pairs of 16-bit units. The I<high
732 surrogates> are the range 0xD800..0xDBFF, and the I<low surrogates>
733 are the range 0xDC00..0xDFFFF. The surrogate encoding is
735 $hi = ($uni - 0x10000) / 0x400 + 0xD800;
736 $lo = ($uni - 0x10000) % 0x400 + 0xDC00;
740 $uni = 0x10000 + ($hi - 0xD8000) * 0x400 + ($lo - 0xDC00);
742 Because of the 16-bitness, UTF-16 is byteorder dependent. UTF-16
743 itself can be used for in-memory computations, but if storage or
744 transfer is required, either UTF-16BE (Big Endian) or UTF-16LE
745 (Little Endian) must be chosen.
747 This introduces another problem: what if you just know that your data
748 is UTF-16, but you don't know which endianness? Byte Order Marks
749 (BOMs) are a solution to this. A special character has been reserved
750 in Unicode to function as a byte order marker: the character with the
751 code point 0xFEFF is the BOM.
753 The trick is that if you read a BOM, you will know the byte order,
754 since if it was written on a big endian platform, you will read the
755 bytes 0xFE 0xFF, but if it was written on a little endian platform,
756 you will read the bytes 0xFF 0xFE. (And if the originating platform
757 was writing in UTF-8, you will read the bytes 0xEF 0xBB 0xBF.)
759 The way this trick works is that the character with the code point
760 0xFFFE is guaranteed not to be a valid Unicode character, so the
761 sequence of bytes 0xFF 0xFE is unambiguously "BOM, represented in
762 little-endian format" and cannot be "0xFFFE, represented in big-endian
765 =item UTF-32, UTF-32BE, UTF32-LE
767 The UTF-32 family is pretty much like the UTF-16 family, expect that
768 the units are 32-bit, and therefore the surrogate scheme is not
769 needed. The BOM signatures will be 0x00 0x00 0xFE 0xFF for BE and
770 0xFF 0xFE 0x00 0x00 for LE.
774 Encodings defined by the ISO 10646 standard. UCS-2 is a 16-bit
775 encoding, UCS-4 is a 32-bit encoding. Unlike UTF-16, UCS-2
776 is not extensible beyond 0xFFFF, because it does not use surrogates.
780 A seven-bit safe (non-eight-bit) encoding, useful if the
781 transport/storage is not eight-bit safe. Defined by RFC 2152.
783 =head2 Security Implications of Malformed UTF-8
785 Unfortunately, the specification of UTF-8 leaves some room for
786 interpretation of how many bytes of encoded output one should generate
787 from one input Unicode character. Strictly speaking, one is supposed
788 to always generate the shortest possible sequence of UTF-8 bytes,
789 because otherwise there is potential for input buffer overflow at the
790 receiving end of a UTF-8 connection. Perl always generates the shortest
791 length UTF-8, and with warnings on (C<-w> or C<use warnings;>) Perl will
792 warn about non-shortest length UTF-8 (and other malformations, too,
793 such as the surrogates, which are not real character code points.)
795 =head2 Unicode in Perl on EBCDIC
797 The way Unicode is handled on EBCDIC platforms is still rather
798 experimental. On such a platform, references to UTF-8 encoding in this
799 document and elsewhere should be read as meaning UTF-EBCDIC as
800 specified in Unicode Technical Report 16 unless ASCII vs EBCDIC issues
801 are specifically discussed. There is no C<utfebcdic> pragma or
802 ":utfebcdic" layer, rather, "utf8" and ":utf8" are re-used to mean
803 the platform's "natural" 8-bit encoding of Unicode. See L<perlebcdic>
804 for more discussion of the issues.
810 L<perluniintro>, L<encoding>, L<Encode>, L<open>, L<utf8>, L<bytes>,
811 L<perlretut>, L<perlvar/"${^WIDE_SYSTEM_CALLS}">