1Encode::Unicode(3pm) Perl Programmers Reference Guide Encode::Unicode(3pm)
2
3
4
6 Encode::Unicode -- Various Unicode Transformation Formats
7
9 use Encode qw/encode decode/;
10 $ucs2 = encode("UCS-2BE", $utf8);
11 $utf8 = decode("UCS-2BE", $ucs2);
12
14 This module implements all Character Encoding Schemes of Unicode that
15 are officially documented by Unicode Consortium (except, of course, for
16 UTF-8, which is a native format in perl).
17
18 <http://www.unicode.org/glossary/> says:
19 Character Encoding Scheme A character encoding form plus byte
20 serialization. There are Seven character encoding schemes in
21 Unicode: UTF-8, UTF-16, UTF-16BE, UTF-16LE, UTF-32 (UCS-4),
22 UTF-32BE (UCS-4BE) and UTF-32LE (UCS-4LE), and UTF-7.
23
24 Since UTF-7 is a 7-bit (re)encoded version of UTF-16BE, It is not
25 part of Unicode's Character Encoding Scheme. It is separately
26 implemented in Encode::Unicode::UTF7. For details see
27 Encode::Unicode::UTF7.
28
29 Quick Reference
30 Decodes from ord(N) Encodes chr(N) to...
31 octet/char BOM S.P d800-dfff ord > 0xffff \x{1abcd} ==
32 ---------------+-----------------+------------------------------
33 UCS-2BE 2 N N is bogus Not Available
34 UCS-2LE 2 N N bogus Not Available
35 UTF-16 2/4 Y Y is S.P S.P BE/LE
36 UTF-16BE 2/4 N Y S.P S.P 0xd82a,0xdfcd
37 UTF-16LE 2/4 N Y S.P S.P 0x2ad8,0xcddf
38 UTF-32 4 Y - is bogus As is BE/LE
39 UTF-32BE 4 N - bogus As is 0x0001abcd
40 UTF-32LE 4 N - bogus As is 0xcdab0100
41 UTF-8 1-4 - - bogus >= 4 octets \xf0\x9a\af\8d
42 ---------------+-----------------+------------------------------
43
45 You can categorize these CES by 3 criteria: size of each character,
46 endianness, and Byte Order Mark.
47
48 by size
49 UCS-2 is a fixed-length encoding with each character taking 16 bits.
50 It does not support surrogate pairs. When a surrogate pair is
51 encountered during decode(), its place is filled with \x{FFFD} if CHECK
52 is 0, or the routine croaks if CHECK is 1. When a character whose ord
53 value is larger than 0xFFFF is encountered, its place is filled with
54 \x{FFFD} if CHECK is 0, or the routine croaks if CHECK is 1.
55
56 UTF-16 is almost the same as UCS-2 but it supports surrogate pairs.
57 When it encounters a high surrogate (0xD800-0xDBFF), it fetches the
58 following low surrogate (0xDC00-0xDFFF) and "desurrogate"s them to form
59 a character. Bogus surrogates result in death. When \x{10000} or
60 above is encountered during encode(), it "ensurrogate"s them and pushes
61 the surrogate pair to the output stream.
62
63 UTF-32 (UCS-4) is a fixed-length encoding with each character taking 32
64 bits. Since it is 32-bit, there is no need for surrogate pairs.
65
66 by endianness
67 The first (and now failed) goal of Unicode was to map all character
68 repertoires into a fixed-length integer so that programmers are happy.
69 Since each character is either a short or long in C, you have to pay
70 attention to the endianness of each platform when you pass data to one
71 another.
72
73 Anything marked as BE is Big Endian (or network byte order) and LE is
74 Little Endian (aka VAX byte order). For anything not marked either BE
75 or LE, a character called Byte Order Mark (BOM) indicating the
76 endianness is prepended to the string.
77
78 CAVEAT: Though BOM in utf8 (\xEF\xBB\xBF) is valid, it is meaningless
79 and as of this writing Encode suite just leave it as is (\x{FeFF}).
80
81 BOM as integer when fetched in network byte order
82 16 32 bits/char
83 -------------------------
84 BE 0xFeFF 0x0000FeFF
85 LE 0xFFFe 0xFFFe0000
86 -------------------------
87
88 This modules handles the BOM as follows.
89
90 · When BE or LE is explicitly stated as the name of encoding, BOM is
91 simply treated as a normal character (ZERO WIDTH NO-BREAK SPACE).
92
93 · When BE or LE is omitted during decode(), it checks if BOM is at
94 the beginning of the string; if one is found, the endianness is set
95 to what the BOM says. If no BOM is found, the routine dies.
96
97 · When BE or LE is omitted during encode(), it returns a BE-encoded
98 string with BOM prepended. So when you want to encode a whole text
99 file, make sure you encode() the whole text at once, not line by
100 line or each line, not file, will have a BOM prepended.
101
102 · "UCS-2" is an exception. Unlike others, this is an alias of
103 UCS-2BE. UCS-2 is already registered by IANA and others that way.
104
106 To say the least, surrogate pairs were the biggest mistake of the
107 Unicode Consortium. But according to the late Douglas Adams in The
108 Hitchhiker's Guide to the Galaxy Trilogy, "In the beginning the
109 Universe was created. This has made a lot of people very angry and been
110 widely regarded as a bad move". Their mistake was not of this
111 magnitude so let's forgive them.
112
113 (I don't dare make any comparison with Unicode Consortium and the
114 Vogons here ;) Or, comparing Encode to Babel Fish is completely
115 appropriate -- if you can only stick this into your ear :)
116
117 Surrogate pairs were born when the Unicode Consortium finally admitted
118 that 16 bits were not big enough to hold all the world's character
119 repertoires. But they already made UCS-2 16-bit. What do we do?
120
121 Back then, the range 0xD800-0xDFFF was not allocated. Let's split that
122 range in half and use the first half to represent the "upper half of a
123 character" and the second half to represent the "lower half of a
124 character". That way, you can represent 1024 * 1024 = 1048576 more
125 characters. Now we can store character ranges up to \x{10ffff} even
126 with 16-bit encodings. This pair of half-character is now called a
127 surrogate pair and UTF-16 is the name of the encoding that embraces
128 them.
129
130 Here is a formula to ensurrogate a Unicode character \x{10000} and
131 above;
132
133 $hi = ($uni - 0x10000) / 0x400 + 0xD800;
134 $lo = ($uni - 0x10000) % 0x400 + 0xDC00;
135
136 And to desurrogate;
137
138 $uni = 0x10000 + ($hi - 0xD800) * 0x400 + ($lo - 0xDC00);
139
140 Note this move has made \x{D800}-\x{DFFF} into a forbidden zone but
141 perl does not prohibit the use of characters within this range. To
142 perl, every one of \x{0000_0000} up to \x{ffff_ffff} (*) is a
143 character.
144
145 (*) or \x{ffff_ffff_ffff_ffff} if your perl is compiled with 64-bit
146 integer support!
147
149 Unlike most encodings which accept various ways to handle errors,
150 Unicode encodings simply croaks.
151
152 % perl -MEncode -e'$_ = "\xfe\xff\xd8\xd9\xda\xdb\0\n"' \
153 -e'Encode::from_to($_, "utf16","shift_jis", 0); print'
154 UTF-16:Malformed LO surrogate d8d9 at /path/to/Encode.pm line 184.
155 % perl -MEncode -e'$a = "BOM missing"' \
156 -e' Encode::from_to($a, "utf16", "shift_jis", 0); print'
157 UTF-16:Unrecognised BOM 424f at /path/to/Encode.pm line 184.
158
159 Unlike other encodings where mappings are not one-to-one against
160 Unicode, UTFs are supposed to map 100% against one another. So Encode
161 is more strict on UTFs.
162
163 Consider that "division by zero" of Encode :)
164
166 Encode, Encode::Unicode::UTF7, <http://www.unicode.org/glossary/>,
167 <http://www.unicode.org/unicode/faq/utf_bom.html>,
168
169 RFC 2781 <http://rfc.net/rfc2781.html>,
170
171 The whole Unicode standard
172 <http://www.unicode.org/unicode/uni2book/u2.html>
173
174 Ch. 15, pp. 403 of "Programming Perl (3rd Edition)" by Larry Wall, Tom
175 Christiansen, Jon Orwant; O'Reilly & Associates; ISBN 0-596-00027-8
176
177
178
179perl v5.10.1 2009-04-14 Encode::Unicode(3pm)