| 1 | package Encode::Unicode; |
| 2 | |
| 3 | use strict; |
| 4 | use warnings; |
| 5 | no warnings 'redefine'; |
| 6 | |
| 7 | our $VERSION = do { my @r = (q$Revision: 2.2 $ =~ /\d+/g); sprintf "%d."."%02d" x $#r, @r }; |
| 8 | |
| 9 | use XSLoader; |
| 10 | XSLoader::load(__PACKAGE__,$VERSION); |
| 11 | |
| 12 | # |
| 13 | # Object Generator 8 transcoders all at once! |
| 14 | # |
| 15 | |
| 16 | require Encode; |
| 17 | |
| 18 | our %BOM_Unknown = map {$_ => 1} qw(UTF-16 UTF-32); |
| 19 | |
| 20 | for my $name (qw(UTF-16 UTF-16BE UTF-16LE |
| 21 | UTF-32 UTF-32BE UTF-32LE |
| 22 | UCS-2BE UCS-2LE)) |
| 23 | { |
| 24 | my ($size, $endian, $ucs2, $mask); |
| 25 | $name =~ /^(\w+)-(\d+)(\w*)$/o; |
| 26 | if ($ucs2 = ($1 eq 'UCS')){ |
| 27 | $size = 2; |
| 28 | }else{ |
| 29 | $size = $2/8; |
| 30 | } |
| 31 | $endian = ($3 eq 'BE') ? 'n' : ($3 eq 'LE') ? 'v' : '' ; |
| 32 | $size == 4 and $endian = uc($endian); |
| 33 | |
| 34 | $Encode::Encoding{$name} = |
| 35 | bless { |
| 36 | Name => $name, |
| 37 | size => $size, |
| 38 | endian => $endian, |
| 39 | ucs2 => $ucs2, |
| 40 | } => __PACKAGE__; |
| 41 | } |
| 42 | |
| 43 | use base qw(Encode::Encoding); |
| 44 | |
| 45 | sub renew { |
| 46 | my $self = shift; |
| 47 | $BOM_Unknown{$self->name} or return $self; |
| 48 | my $clone = bless { %$self } => ref($self); |
| 49 | $clone->{renewed}++; # so the caller knows it is renewed. |
| 50 | return $clone; |
| 51 | } |
| 52 | |
| 53 | # There used to be a perl implemntation of (en|de)code but with |
| 54 | # XS version is ripe, perl version is zapped for optimal speed |
| 55 | |
| 56 | *decode = \&decode_xs; |
| 57 | *encode = \&encode_xs; |
| 58 | |
| 59 | 1; |
| 60 | __END__ |
| 61 | |
| 62 | =head1 NAME |
| 63 | |
| 64 | Encode::Unicode -- Various Unicode Transformation Formats |
| 65 | |
| 66 | =cut |
| 67 | |
| 68 | =head1 SYNOPSIS |
| 69 | |
| 70 | use Encode qw/encode decode/; |
| 71 | $ucs2 = encode("UCS-2BE", $utf8); |
| 72 | $utf8 = decode("UCS-2BE", $ucs2); |
| 73 | |
| 74 | =head1 ABSTRACT |
| 75 | |
| 76 | This module implements all Character Encoding Schemes of Unicode that |
| 77 | are officially documented by Unicode Consortium (except, of course, |
| 78 | for UTF-8, which is a native format in perl). |
| 79 | |
| 80 | =over 4 |
| 81 | |
| 82 | =item L<http://www.unicode.org/glossary/> says: |
| 83 | |
| 84 | I<Character Encoding Scheme> A character encoding form plus byte |
| 85 | serialization. There are Seven character encoding schemes in Unicode: |
| 86 | UTF-8, UTF-16, UTF-16BE, UTF-16LE, UTF-32 (UCS-4), UTF-32BE (UCS-4BE) and |
| 87 | UTF-32LE (UCS-4LE), and UTF-7. |
| 88 | |
| 89 | Since UTF-7 is a 7-bit (re)encoded version of UTF-16BE, It is not part of |
| 90 | Unicode's Character Encoding Scheme. It is separately implemented in |
| 91 | Encode::Unicode::UTF7. For details see L<Encode::Unicode::UTF7>. |
| 92 | |
| 93 | =item Quick Reference |
| 94 | |
| 95 | Decodes from ord(N) Encodes chr(N) to... |
| 96 | octet/char BOM S.P d800-dfff ord > 0xffff \x{1abcd} == |
| 97 | ---------------+-----------------+------------------------------ |
| 98 | UCS-2BE 2 N N is bogus Not Available |
| 99 | UCS-2LE 2 N N bogus Not Available |
| 100 | UTF-16 2/4 Y Y is S.P S.P BE/LE |
| 101 | UTF-16BE 2/4 N Y S.P S.P 0xd82a,0xdfcd |
| 102 | UTF-16LE 2 N Y S.P S.P 0x2ad8,0xcddf |
| 103 | UTF-32 4 Y - is bogus As is BE/LE |
| 104 | UTF-32BE 4 N - bogus As is 0x0001abcd |
| 105 | UTF-32LE 4 N - bogus As is 0xcdab0100 |
| 106 | UTF-8 1-4 - - bogus >= 4 octets \xf0\x9a\af\8d |
| 107 | ---------------+-----------------+------------------------------ |
| 108 | |
| 109 | =back |
| 110 | |
| 111 | =head1 Size, Endianness, and BOM |
| 112 | |
| 113 | You can categorize these CES by 3 criteria: size of each character, |
| 114 | endianness, and Byte Order Mark. |
| 115 | |
| 116 | =head2 by size |
| 117 | |
| 118 | UCS-2 is a fixed-length encoding with each character taking 16 bits. |
| 119 | It B<does not> support I<surrogate pairs>. When a surrogate pair |
| 120 | is encountered during decode(), its place is filled with \x{FFFD} |
| 121 | if I<CHECK> is 0, or the routine croaks if I<CHECK> is 1. When a |
| 122 | character whose ord value is larger than 0xFFFF is encountered, |
| 123 | its place is filled with \x{FFFD} if I<CHECK> is 0, or the routine |
| 124 | croaks if I<CHECK> is 1. |
| 125 | |
| 126 | UTF-16 is almost the same as UCS-2 but it supports I<surrogate pairs>. |
| 127 | When it encounters a high surrogate (0xD800-0xDBFF), it fetches the |
| 128 | following low surrogate (0xDC00-0xDFFF) and C<desurrogate>s them to |
| 129 | form a character. Bogus surrogates result in death. When \x{10000} |
| 130 | or above is encountered during encode(), it C<ensurrogate>s them and |
| 131 | pushes the surrogate pair to the output stream. |
| 132 | |
| 133 | UTF-32 (UCS-4) is a fixed-length encoding with each character taking 32 bits. |
| 134 | Since it is 32-bit, there is no need for I<surrogate pairs>. |
| 135 | |
| 136 | =head2 by endianness |
| 137 | |
| 138 | The first (and now failed) goal of Unicode was to map all character |
| 139 | repertoires into a fixed-length integer so that programmers are happy. |
| 140 | Since each character is either a I<short> or I<long> in C, you have to |
| 141 | pay attention to the endianness of each platform when you pass data |
| 142 | to one another. |
| 143 | |
| 144 | Anything marked as BE is Big Endian (or network byte order) and LE is |
| 145 | Little Endian (aka VAX byte order). For anything not marked either |
| 146 | BE or LE, a character called Byte Order Mark (BOM) indicating the |
| 147 | endianness is prepended to the string. |
| 148 | |
| 149 | CAVEAT: Though BOM in utf8 (\xEF\xBB\xBF) is valid, it is meaningless |
| 150 | and as of this writing Encode suite just leave it as is (\x{FeFF}). |
| 151 | |
| 152 | =over 4 |
| 153 | |
| 154 | =item BOM as integer when fetched in network byte order |
| 155 | |
| 156 | 16 32 bits/char |
| 157 | ------------------------- |
| 158 | BE 0xFeFF 0x0000FeFF |
| 159 | LE 0xFFeF 0xFFFe0000 |
| 160 | ------------------------- |
| 161 | |
| 162 | =back |
| 163 | |
| 164 | This modules handles the BOM as follows. |
| 165 | |
| 166 | =over 4 |
| 167 | |
| 168 | =item * |
| 169 | |
| 170 | When BE or LE is explicitly stated as the name of encoding, BOM is |
| 171 | simply treated as a normal character (ZERO WIDTH NO-BREAK SPACE). |
| 172 | |
| 173 | =item * |
| 174 | |
| 175 | When BE or LE is omitted during decode(), it checks if BOM is at the |
| 176 | beginning of the string; if one is found, the endianness is set to |
| 177 | what the BOM says. If no BOM is found, the routine dies. |
| 178 | |
| 179 | =item * |
| 180 | |
| 181 | When BE or LE is omitted during encode(), it returns a BE-encoded |
| 182 | string with BOM prepended. So when you want to encode a whole text |
| 183 | file, make sure you encode() the whole text at once, not line by line |
| 184 | or each line, not file, will have a BOM prepended. |
| 185 | |
| 186 | =item * |
| 187 | |
| 188 | C<UCS-2> is an exception. Unlike others, this is an alias of UCS-2BE. |
| 189 | UCS-2 is already registered by IANA and others that way. |
| 190 | |
| 191 | =back |
| 192 | |
| 193 | =head1 Surrogate Pairs |
| 194 | |
| 195 | To say the least, surrogate pairs were the biggest mistake of the |
| 196 | Unicode Consortium. But according to the late Douglas Adams in I<The |
| 197 | Hitchhiker's Guide to the Galaxy> Trilogy, C<In the beginning the |
| 198 | Universe was created. This has made a lot of people very angry and |
| 199 | been widely regarded as a bad move>. Their mistake was not of this |
| 200 | magnitude so let's forgive them. |
| 201 | |
| 202 | (I don't dare make any comparison with Unicode Consortium and the |
| 203 | Vogons here ;) Or, comparing Encode to Babel Fish is completely |
| 204 | appropriate -- if you can only stick this into your ear :) |
| 205 | |
| 206 | Surrogate pairs were born when the Unicode Consortium finally |
| 207 | admitted that 16 bits were not big enough to hold all the world's |
| 208 | character repertoires. But they already made UCS-2 16-bit. What |
| 209 | do we do? |
| 210 | |
| 211 | Back then, the range 0xD800-0xDFFF was not allocated. Let's split |
| 212 | that range in half and use the first half to represent the C<upper |
| 213 | half of a character> and the second half to represent the C<lower |
| 214 | half of a character>. That way, you can represent 1024 * 1024 = |
| 215 | 1048576 more characters. Now we can store character ranges up to |
| 216 | \x{10ffff} even with 16-bit encodings. This pair of half-character is |
| 217 | now called a I<surrogate pair> and UTF-16 is the name of the encoding |
| 218 | that embraces them. |
| 219 | |
| 220 | Here is a formula to ensurrogate a Unicode character \x{10000} and |
| 221 | above; |
| 222 | |
| 223 | $hi = ($uni - 0x10000) / 0x400 + 0xD800; |
| 224 | $lo = ($uni - 0x10000) % 0x400 + 0xDC00; |
| 225 | |
| 226 | And to desurrogate; |
| 227 | |
| 228 | $uni = 0x10000 + ($hi - 0xD800) * 0x400 + ($lo - 0xDC00); |
| 229 | |
| 230 | Note this move has made \x{D800}-\x{DFFF} into a forbidden zone but |
| 231 | perl does not prohibit the use of characters within this range. To perl, |
| 232 | every one of \x{0000_0000} up to \x{ffff_ffff} (*) is I<a character>. |
| 233 | |
| 234 | (*) or \x{ffff_ffff_ffff_ffff} if your perl is compiled with 64-bit |
| 235 | integer support! |
| 236 | |
| 237 | =head1 Error Checking |
| 238 | |
| 239 | Unlike most encodings which accept various ways to handle errors, |
| 240 | Unicode encodings simply croaks. |
| 241 | |
| 242 | % perl -MEncode -e '$_ = "\xfe\xff\xd8\xd9\xda\xdb\0\n"' \ |
| 243 | -e 'Encode::from_to($_, "utf16","shift_jis", 0); print' |
| 244 | UTF-16:Malformed LO surrogate d8d9 at /path/to/Encode.pm line 184. |
| 245 | % perl -MEncode -e '$a = "BOM missing"' \ |
| 246 | -e ' Encode::from_to($a, "utf16", "shift_jis", 0); print' |
| 247 | UTF-16:Unrecognised BOM 424f at /path/to/Encode.pm line 184. |
| 248 | |
| 249 | Unlike other encodings where mappings are not one-to-one against |
| 250 | Unicode, UTFs are supposed to map 100% against one another. So Encode |
| 251 | is more strict on UTFs. |
| 252 | |
| 253 | Consider that "division by zero" of Encode :) |
| 254 | |
| 255 | =head1 SEE ALSO |
| 256 | |
| 257 | L<Encode>, L<Encode::Unicode::UTF7>, L<http://www.unicode.org/glossary/>, |
| 258 | L<http://www.unicode.org/unicode/faq/utf_bom.html>, |
| 259 | |
| 260 | RFC 2781 L<http://rfc.net/rfc2781.html>, |
| 261 | |
| 262 | The whole Unicode standard L<http://www.unicode.org/unicode/uni2book/u2.html> |
| 263 | |
| 264 | Ch. 15, pp. 403 of C<Programming Perl (3rd Edition)> |
| 265 | by Larry Wall, Tom Christiansen, Jon Orwant; |
| 266 | O'Reilly & Associates; ISBN 0-596-00027-8 |
| 267 | |
| 268 | =cut |