Commit | Line | Data |
2c674647 |
1 | package Encode; |
51ef4e11 |
2 | use strict; |
0e567a6c |
3 | our $VERSION = '0.40'; |
2c674647 |
4 | |
5 | require DynaLoader; |
6 | require Exporter; |
7 | |
51ef4e11 |
8 | our @ISA = qw(Exporter DynaLoader); |
2c674647 |
9 | |
4411f3b6 |
10 | # Public, encouraged API is exported by default |
51ef4e11 |
11 | our @EXPORT = qw ( |
4411f3b6 |
12 | encode |
13 | decode |
14 | encode_utf8 |
15 | decode_utf8 |
16 | find_encoding |
51ef4e11 |
17 | encodings |
4411f3b6 |
18 | ); |
19 | |
51ef4e11 |
20 | our @EXPORT_OK = |
2c674647 |
21 | qw( |
51ef4e11 |
22 | define_encoding |
23 | define_alias |
2c674647 |
24 | from_to |
25 | is_utf8 |
4411f3b6 |
26 | is_8bit |
27 | is_16bit |
a12c0f56 |
28 | utf8_upgrade |
29 | utf8_downgrade |
4411f3b6 |
30 | _utf8_on |
31 | _utf8_off |
2c674647 |
32 | ); |
33 | |
34 | bootstrap Encode (); |
35 | |
4411f3b6 |
36 | # Documentation moved after __END__ for speed - NI-S |
2c674647 |
37 | |
bf230f3d |
38 | use Carp; |
39 | |
51ef4e11 |
40 | # Make a %encoding package variable to allow a certain amount of cheating |
41 | our %encoding; |
42 | my @alias; # ordered matching list |
43 | my %alias; # cached known aliases |
f7ac3676 |
44 | |
6d6a7c8d |
45 | # 0 1 2 3 4 5 6 7 8 9 10 |
46 | our @latin2iso_num = ( 0, 1, 2, 3, 4, 9, 10, 13, 14, 15, 16 ); |
47 | |
f7ac3676 |
48 | our %winlatin2cp = ( |
49 | 'Latin1' => 1252, |
50 | 'Latin2' => 1250, |
51 | 'Cyrillic' => 1251, |
f7ac3676 |
52 | 'Greek' => 1253, |
53 | 'Turkish' => 1254, |
54 | 'Hebrew' => 1255, |
55 | 'Arabic' => 1256, |
56 | 'Baltic' => 1257, |
57 | 'Vietnamese' => 1258, |
58 | ); |
5345d506 |
59 | |
2b217bf7 |
60 | our %external_tables = |
61 | ( |
62 | 'euc-cn' => 'Encode/CN.pm', |
63 | gb2312 => 'Encode/CN.pm', |
64 | gb12345 => 'Encode/CN.pm', |
65 | gbk => 'Encode/CN.pm', |
66 | cp936 => 'Encode/CN.pm', |
67 | 'iso-ir-165' => 'Encode/CN.pm', |
68 | 'euc-jp' => 'Encode/JP.pm', |
69 | shiftjis => 'Encode/JP.pm', |
70 | macjapan => 'Encode/JP.pm', |
71 | cp932 => 'Encode/JP.pm', |
72 | 'euc-kr' => 'Encode/KR.pm', |
73 | ksc5601 => 'Encode/KR.pm', |
74 | cp949 => 'Encode/KR.pm', |
75 | big5 => 'Encode/TW.pm', |
76 | 'big5-hkscs' => 'Encode/TW.pm', |
77 | cp950 => 'Encode/TW.pm', |
78 | gb18030 => 'Encode/HanExtra.pm', |
79 | big5plus => 'Encode/HanExtra.pm', |
80 | 'euc-tw' => 'Encode/HanExtra.pm', |
81 | ); |
d1ed7747 |
82 | |
656753f8 |
83 | sub encodings |
84 | { |
85 | my ($class) = @_; |
40a073c6 |
86 | return |
87 | map { $_->[0] } |
88 | sort { $a->[1] cmp $b->[1] } |
89 | map { [$_, lc $_] } |
90 | grep { $_ ne 'Internal' } |
91 | keys %encoding; |
51ef4e11 |
92 | } |
93 | |
94 | sub findAlias |
95 | { |
18586f54 |
96 | my $class = shift; |
97 | local $_ = shift; |
98 | # print "# findAlias $_\n"; |
99 | unless (exists $alias{$_}) |
656753f8 |
100 | { |
18586f54 |
101 | for (my $i=0; $i < @alias; $i += 2) |
102 | { |
103 | my $alias = $alias[$i]; |
104 | my $val = $alias[$i+1]; |
105 | my $new; |
106 | if (ref($alias) eq 'Regexp' && $_ =~ $alias) |
107 | { |
108 | $new = eval $val; |
109 | } |
110 | elsif (ref($alias) eq 'CODE') |
111 | { |
112 | $new = &{$alias}($val) |
113 | } |
114 | elsif (lc($_) eq lc($alias)) |
115 | { |
116 | $new = $val; |
117 | } |
118 | if (defined($new)) |
119 | { |
120 | next if $new eq $_; # avoid (direct) recursion on bugs |
121 | my $enc = (ref($new)) ? $new : find_encoding($new); |
122 | if ($enc) |
123 | { |
124 | $alias{$_} = $enc; |
125 | last; |
126 | } |
127 | } |
128 | } |
656753f8 |
129 | } |
18586f54 |
130 | return $alias{$_}; |
5345d506 |
131 | } |
132 | |
51ef4e11 |
133 | sub define_alias |
5345d506 |
134 | { |
18586f54 |
135 | while (@_) |
136 | { |
137 | my ($alias,$name) = splice(@_,0,2); |
138 | push(@alias, $alias => $name); |
139 | } |
51ef4e11 |
140 | } |
141 | |
016cb72c |
142 | # Allow variants of iso-8859-1 etc. |
d6089a2a |
143 | define_alias( qr/^iso[-_]?(\d+)[-_](\d+)$/i => '"iso-$1-$2"' ); |
016cb72c |
144 | |
7faf300d |
145 | # At least HP-UX has these. |
146 | define_alias( qr/^iso8859(\d+)$/i => '"iso-8859-$1"' ); |
147 | |
f7ac3676 |
148 | # More HP stuff. |
149 | define_alias( qr/^(?:hp-)?(arabic|greek|hebrew|kana|roman|thai|turkish)8$/i => '"${1}8"' ); |
150 | |
0b3236bb |
151 | # The Official name of ASCII. |
8a361256 |
152 | define_alias( qr/^ANSI[-_]?X3\.4[-_]?1968$/i => '"ascii"' ); |
153 | |
58d53262 |
154 | # This is a font issue, not an encoding issue. |
155 | # (The currency symbol of the Latin 1 upper half |
156 | # has been redefined as the euro symbol.) |
157 | define_alias( qr/^(.+)\@euro$/i => '"$1"' ); |
158 | |
016cb72c |
159 | # Allow latin-1 style names as well |
7faf300d |
160 | define_alias( qr/^(?:iso[-_]?)?latin[-_]?(\d+)$/i => '"iso-8859-$latin2iso_num[$1]"' ); |
016cb72c |
161 | |
f7ac3676 |
162 | # Allow winlatin1 style names as well |
cf91068f |
163 | define_alias( qr/^win(latin[12]|cyrillic|baltic|greek|turkish|hebrew|arabic|baltic|vietnamese)$/i => '"cp$winlatin2cp{\u$1}"' ); |
f7ac3676 |
164 | |
016cb72c |
165 | # Common names for non-latin prefered MIME names |
166 | define_alias( 'ascii' => 'US-ascii', |
167 | 'cyrillic' => 'iso-8859-5', |
168 | 'arabic' => 'iso-8859-6', |
169 | 'greek' => 'iso-8859-7', |
f7ac3676 |
170 | 'hebrew' => 'iso-8859-8', |
171 | 'thai' => 'iso-8859-11', |
172 | 'tis620' => 'iso-8859-11', |
173 | ); |
016cb72c |
174 | |
7faf300d |
175 | # At least AIX has IBM-NNN (surprisingly...) instead of cpNNN. |
1853dd5f |
176 | # And Microsoft has their own naming (again, surprisingly). |
177 | define_alias( qr/^(?:ibm|ms)[-_]?(\d\d\d\d?)$/i => '"cp$1"'); |
178 | |
179 | # Sometimes seen with a leading zero. |
180 | define_alias( qr/^cp037$/i => '"cp37"'); |
181 | |
182 | # Ououououou. |
183 | define_alias( qr/^macRomanian$/i => '"macRumanian"'); |
7faf300d |
184 | |
58d53262 |
185 | # Standardize on the dashed versions. |
186 | define_alias( qr/^utf8$/i => 'utf-8' ); |
7faf300d |
187 | define_alias( qr/^koi8r$/i => 'koi8-r' ); |
f7ac3676 |
188 | define_alias( qr/^koi8u$/i => 'koi8-u' ); |
189 | |
1853dd5f |
190 | # Seen in some Linuxes. |
191 | define_alias( qr/^ujis$/i => 'euc-jp' ); |
192 | |
b2729934 |
193 | # CP936 doesn't have vendor-addon for GBK, so they're identical. |
194 | define_alias( qr/^gbk$/i => '"cp936"'); |
195 | |
f7ac3676 |
196 | # TODO: HP-UX '8' encodings arabic8 greek8 hebrew8 kana8 thai8 turkish8 |
197 | # TODO: HP-UX '15' encodings japanese15 korean15 roi15 |
198 | # TODO: Cyrillic encoding ISO-IR-111 (useful?) |
f7ac3676 |
199 | # TODO: Armenian encoding ARMSCII-8 |
200 | # TODO: Hebrew encoding ISO-8859-8-1 |
201 | # TODO: Thai encoding TCVN |
202 | # TODO: Korean encoding Johab |
56a543c5 |
203 | # TODO: Vietnamese encodings VPS |
f7ac3676 |
204 | # TODO: Japanese encoding JIS (not the same as SJIS) |
205 | # TODO: Mac Asian+African encodings: Arabic Armenian Bengali Burmese |
206 | # ChineseSimp ChineseTrad Devanagari Ethiopic ExtArabic |
207 | # Farsi Georgian Gujarati Gurmukhi Hebrew Japanese |
208 | # Kannada Khmer Korean Laotian Malayalam Mongolian |
209 | # Oriya Sinhalese Symbol Tamil Telugu Tibetan Vietnamese |
18586f54 |
210 | |
1853dd5f |
211 | # Map white space and _ to '-' |
016cb72c |
212 | define_alias( qr/^(\S+)[\s_]+(.*)$/i => '"$1-$2"' ); |
213 | |
51ef4e11 |
214 | sub define_encoding |
215 | { |
18586f54 |
216 | my $obj = shift; |
217 | my $name = shift; |
218 | $encoding{$name} = $obj; |
219 | my $lc = lc($name); |
220 | define_alias($lc => $obj) unless $lc eq $name; |
221 | while (@_) |
222 | { |
223 | my $alias = shift; |
224 | define_alias($alias,$obj); |
225 | } |
226 | return $obj; |
656753f8 |
227 | } |
228 | |
656753f8 |
229 | sub getEncoding |
230 | { |
dd9703c9 |
231 | my ($class,$name,$skip_external) = @_; |
18586f54 |
232 | my $enc; |
233 | if (ref($name) && $name->can('new_sequence')) |
234 | { |
235 | return $name; |
236 | } |
237 | my $lc = lc $name; |
238 | if (exists $encoding{$name}) |
239 | { |
240 | return $encoding{$name}; |
241 | } |
242 | if (exists $encoding{$lc}) |
243 | { |
244 | return $encoding{$lc}; |
245 | } |
c50d192e |
246 | |
247 | my $oc = $class->findAlias($name); |
248 | return $oc if defined $oc; |
249 | |
250 | $oc = $class->findAlias($lc) if $lc ne $name; |
251 | return $oc if defined $oc; |
252 | |
dd9703c9 |
253 | if (!$skip_external and exists $external_tables{$lc}) |
d1ed7747 |
254 | { |
255 | require $external_tables{$lc}; |
256 | return $encoding{$name} if exists $encoding{$name}; |
257 | } |
18586f54 |
258 | |
18586f54 |
259 | return; |
656753f8 |
260 | } |
261 | |
4411f3b6 |
262 | sub find_encoding |
263 | { |
dd9703c9 |
264 | my ($name,$skip_external) = @_; |
265 | return __PACKAGE__->getEncoding($name,$skip_external); |
4411f3b6 |
266 | } |
267 | |
268 | sub encode |
269 | { |
18586f54 |
270 | my ($name,$string,$check) = @_; |
271 | my $enc = find_encoding($name); |
272 | croak("Unknown encoding '$name'") unless defined $enc; |
273 | my $octets = $enc->encode($string,$check); |
274 | return undef if ($check && length($string)); |
275 | return $octets; |
4411f3b6 |
276 | } |
277 | |
278 | sub decode |
279 | { |
18586f54 |
280 | my ($name,$octets,$check) = @_; |
281 | my $enc = find_encoding($name); |
282 | croak("Unknown encoding '$name'") unless defined $enc; |
283 | my $string = $enc->decode($octets,$check); |
284 | $_[1] = $octets if $check; |
285 | return $string; |
4411f3b6 |
286 | } |
287 | |
288 | sub from_to |
289 | { |
18586f54 |
290 | my ($string,$from,$to,$check) = @_; |
291 | my $f = find_encoding($from); |
292 | croak("Unknown encoding '$from'") unless defined $f; |
293 | my $t = find_encoding($to); |
294 | croak("Unknown encoding '$to'") unless defined $t; |
295 | my $uni = $f->decode($string,$check); |
296 | return undef if ($check && length($string)); |
297 | $string = $t->encode($uni,$check); |
298 | return undef if ($check && length($uni)); |
299 | return length($_[0] = $string); |
4411f3b6 |
300 | } |
301 | |
302 | sub encode_utf8 |
303 | { |
18586f54 |
304 | my ($str) = @_; |
305 | utf8::encode($str); |
306 | return $str; |
4411f3b6 |
307 | } |
308 | |
309 | sub decode_utf8 |
310 | { |
18586f54 |
311 | my ($str) = @_; |
312 | return undef unless utf8::decode($str); |
313 | return $str; |
5ad8ef52 |
314 | } |
315 | |
18586f54 |
316 | require Encode::Encoding; |
317 | require Encode::XS; |
318 | require Encode::Internal; |
319 | require Encode::Unicode; |
320 | require Encode::utf8; |
321 | require Encode::iso10646_1; |
322 | require Encode::ucs2_le; |
4411f3b6 |
323 | |
656753f8 |
324 | 1; |
325 | |
2a936312 |
326 | __END__ |
327 | |
4411f3b6 |
328 | =head1 NAME |
329 | |
330 | Encode - character encodings |
331 | |
332 | =head1 SYNOPSIS |
333 | |
334 | use Encode; |
335 | |
336 | =head1 DESCRIPTION |
337 | |
47bfe92f |
338 | The C<Encode> module provides the interfaces between Perl's strings |
339 | and the rest of the system. Perl strings are sequences of B<characters>. |
4411f3b6 |
340 | |
341 | The repertoire of characters that Perl can represent is at least that |
47bfe92f |
342 | defined by the Unicode Consortium. On most platforms the ordinal |
343 | values of the characters (as returned by C<ord(ch)>) is the "Unicode |
344 | codepoint" for the character (the exceptions are those platforms where |
345 | the legacy encoding is some variant of EBCDIC rather than a super-set |
346 | of ASCII - see L<perlebcdic>). |
4411f3b6 |
347 | |
348 | Traditionaly computer data has been moved around in 8-bit chunks |
349 | often called "bytes". These chunks are also known as "octets" in |
350 | networking standards. Perl is widely used to manipulate data of |
351 | many types - not only strings of characters representing human or |
352 | computer languages but also "binary" data being the machines representation |
353 | of numbers, pixels in an image - or just about anything. |
354 | |
47bfe92f |
355 | When Perl is processing "binary data" the programmer wants Perl to process |
356 | "sequences of bytes". This is not a problem for Perl - as a byte has 256 |
357 | possible values it easily fits in Perl's much larger "logical character". |
4411f3b6 |
358 | |
d1ed7747 |
359 | Due to size concerns, each of B<CJK> (Chinese, Japanese & Korean) modules |
360 | are not loaded in memory until the first time they're used. Although you |
361 | don't have to C<use> the corresponding B<Encode::>(B<TW>|B<CN>|B<JP>|B<KR>) |
362 | modules first, be aware that those encodings will not be in C<%encodings> |
363 | until their module is loaded (either implicitly through using encodings |
364 | contained in the same module, or via an explicit C<use>). |
a67efb5b |
365 | |
4411f3b6 |
366 | =head2 TERMINOLOGY |
367 | |
4ac9195f |
368 | =over 4 |
4411f3b6 |
369 | |
370 | =item * |
371 | |
372 | I<character>: a character in the range 0..(2**32-1) (or more). |
47bfe92f |
373 | (What Perl's strings are made of.) |
4411f3b6 |
374 | |
375 | =item * |
376 | |
377 | I<byte>: a character in the range 0..255 |
47bfe92f |
378 | (A special case of a Perl character.) |
4411f3b6 |
379 | |
380 | =item * |
381 | |
382 | I<octet>: 8 bits of data, with ordinal values 0..255 |
47bfe92f |
383 | (Term for bytes passed to or from a non-Perl context, e.g. disk file.) |
4411f3b6 |
384 | |
385 | =back |
386 | |
387 | The marker [INTERNAL] marks Internal Implementation Details, in |
388 | general meant only for those who think they know what they are doing, |
389 | and such details may change in future releases. |
390 | |
391 | =head1 ENCODINGS |
392 | |
393 | =head2 Characteristics of an Encoding |
394 | |
395 | An encoding has a "repertoire" of characters that it can represent, |
396 | and for each representable character there is at least one sequence of |
397 | octets that represents it. |
398 | |
399 | =head2 Types of Encodings |
400 | |
401 | Encodings can be divided into the following types: |
402 | |
403 | =over 4 |
404 | |
405 | =item * Fixed length 8-bit (or less) encodings. |
406 | |
407 | Each character is a single octet so may have a repertoire of up to |
408 | 256 characters. ASCII and iso-8859-* are typical examples. |
409 | |
410 | =item * Fixed length 16-bit encodings |
411 | |
412 | Each character is two octets so may have a repertoire of up to |
47bfe92f |
413 | 65 536 characters. Unicode's UCS-2 is an example. Also used for |
4411f3b6 |
414 | encodings for East Asian languages. |
415 | |
416 | =item * Fixed length 32-bit encodings. |
417 | |
418 | Not really very "encoded" encodings. The Unicode code points |
419 | are just represented as 4-octet integers. None the less because |
420 | different architectures use different representations of integers |
421 | (so called "endian") there at least two disctinct encodings. |
422 | |
423 | =item * Multi-byte encodings |
424 | |
425 | The number of octets needed to represent a character varies. |
426 | UTF-8 is a particularly complex but regular case of a multi-byte |
427 | encoding. Several East Asian countries use a multi-byte encoding |
428 | where 1-octet is used to cover western roman characters and Asian |
429 | characters get 2-octets. |
430 | (UTF-16 is strictly a multi-byte encoding taking either 2 or 4 octets |
431 | to represent a Unicode code point.) |
432 | |
433 | =item * "Escape" encodings. |
434 | |
435 | These encodings embed "escape sequences" into the octet sequence |
436 | which describe how the following octets are to be interpreted. |
437 | The iso-2022-* family is typical. Following the escape sequence |
438 | octets are encoded by an "embedded" encoding (which will be one |
439 | of the above types) until another escape sequence switches to |
440 | a different "embedded" encoding. |
441 | |
442 | These schemes are very flexible and can handle mixed languages but are |
47bfe92f |
443 | very complex to process (and have state). No escape encodings are |
444 | implemented for Perl yet. |
4411f3b6 |
445 | |
446 | =back |
447 | |
448 | =head2 Specifying Encodings |
449 | |
450 | Encodings can be specified to the API described below in two ways: |
451 | |
452 | =over 4 |
453 | |
454 | =item 1. By name |
455 | |
47bfe92f |
456 | Encoding names are strings with characters taken from a restricted |
457 | repertoire. See L</"Encoding Names">. |
4411f3b6 |
458 | |
459 | =item 2. As an object |
460 | |
2659725b |
461 | Encoding objects are returned by C<find_encoding($name, [$skip_external])>. |
462 | If the second parameter is true, Encode will refrain from loading external |
463 | modules for CJK encodings. |
4411f3b6 |
464 | |
465 | =back |
466 | |
467 | =head2 Encoding Names |
468 | |
469 | Encoding names are case insensitive. White space in names is ignored. |
47bfe92f |
470 | In addition an encoding may have aliases. Each encoding has one |
471 | "canonical" name. The "canonical" name is chosen from the names of |
472 | the encoding by picking the first in the following sequence: |
4411f3b6 |
473 | |
474 | =over 4 |
475 | |
78255929 |
476 | =item * The MIME name as defined in IETF RFCs. |
4411f3b6 |
477 | |
478 | =item * The name in the IANA registry. |
479 | |
d1be9408 |
480 | =item * The name used by the organization that defined it. |
4411f3b6 |
481 | |
482 | =back |
483 | |
484 | Because of all the alias issues, and because in the general case |
485 | encodings have state C<Encode> uses the encoding object internally |
486 | once an operation is in progress. |
487 | |
21938dfa |
488 | As of Perl 5.8.0, at least the following encodings are recognized |
489 | (the => marks aliases): |
490 | |
491 | ASCII |
492 | |
493 | US-ASCII => ASCII |
494 | |
495 | The Unicode: |
496 | |
0b3236bb |
497 | UTF-8 |
21938dfa |
498 | UTF-16 |
499 | UCS-2 |
500 | |
501 | ISO 10646-1 => UCS-2 |
502 | |
503 | The ISO 8859 and KOI: |
504 | |
505 | ISO 8859-1 ISO 8859-6 ISO 8859-11 KOI8-F |
506 | ISO 8859-2 ISO 8859-7 (12 doesn't exist) KOI8-R |
56a543c5 |
507 | ISO 8859-3 ISO 8859-8 ISO 8859-13 KOI8-U |
21938dfa |
508 | ISO 8859-4 ISO 8859-9 ISO 8859-14 |
509 | ISO 8859-5 ISO 8859-10 ISO 8859-15 |
510 | ISO 8859-16 |
511 | |
512 | Latin1 => 8859-1 Latin6 => 8859-10 |
513 | Latin2 => 8859-2 Latin7 => 8859-13 |
0b3236bb |
514 | Latin3 => 8859-3 Latin8 => 8859-14 |
21938dfa |
515 | Latin4 => 8859-4 Latin9 => 8859-15 |
516 | Latin5 => 8859-9 Latin10 => 8859-16 |
517 | |
518 | Cyrillic => 8859-5 |
519 | Arabic => 8859-6 |
520 | Greek => 8859-7 |
521 | Hebrew => 8859-8 |
522 | Thai => 8859-11 |
0b3236bb |
523 | TIS620 => 8859-11 |
21938dfa |
524 | |
525 | The CJKV: Chinese, Japanese, Korean, Vietnamese: |
526 | |
0b3236bb |
527 | ISO 2022 ISO 2022 JP-1 JIS 0201 GB 1988 Big5 EUC-CN |
528 | ISO 2022 CN ISO 2022 JP-2 JIS 0208 GB 2312 HZ EUC-JP |
21938dfa |
529 | ISO 2022 JP ISO 2022 KR JIS 0210 GB 12345 CNS 11643 EUC-JP-0212 |
cb448690 |
530 | Shift-JIS GBK Big5-HKSCS EUC-KR |
531 | VISCII ISO-IR-165 |
532 | |
a67efb5b |
533 | (Due to size concerns, additional Chinese encodings including C<GB 18030>, |
534 | C<EUC-TW> and C<BIG5PLUS> are distributed separately on CPAN, under the name |
535 | L<Encode::HanExtra>.) |
21938dfa |
536 | |
537 | The PC codepages: |
538 | |
539 | CP37 CP852 CP861 CP866 CP949 CP1251 CP1256 |
540 | CP424 CP855 CP862 CP869 CP950 CP1252 CP1257 |
541 | CP737 CP856 CP863 CP874 CP1006 CP1253 CP1258 |
542 | CP775 CP857 CP864 CP932 CP1047 CP1254 |
543 | CP850 CP860 CP865 CP936 CP1250 CP1255 |
544 | |
545 | WinLatin1 => CP1252 |
546 | WinLatin2 => CP1250 |
547 | WinCyrillic => CP1251 |
548 | WinGreek => CP1253 |
549 | WinTurkiskh => CP1254 |
550 | WinHebrew => CP1255 |
551 | WinArabic => CP1256 |
552 | WinBaltic => CP1257 |
553 | WinVietnamese => CP1258 |
554 | |
4a42e14c |
555 | (All the CPI<NNN...> are available also as IBMI<NNN...>.) |
21938dfa |
556 | |
557 | The Mac codepages: |
558 | |
0b3236bb |
559 | MacCentralEuropean MacJapanese |
560 | MacCroatian MacRoman |
1853dd5f |
561 | MacCyrillic MacRomanian |
0b3236bb |
562 | MacDingbats MacSami |
563 | MacGreek MacThai |
564 | MacIcelandic MacTurkish |
565 | MacUkraine |
21938dfa |
566 | |
567 | Miscellaneous: |
568 | |
569 | 7bit-greek IR-197 |
570 | 7bit-kana NeXTstep |
571 | 7bit-latin1 POSIX-BC |
572 | DingBats Roman8 |
573 | GSM 0338 Symbol |
574 | |
4411f3b6 |
575 | =head1 PERL ENCODING API |
576 | |
577 | =head2 Generic Encoding Interface |
578 | |
579 | =over 4 |
580 | |
581 | =item * |
582 | |
583 | $bytes = encode(ENCODING, $string[, CHECK]) |
584 | |
47bfe92f |
585 | Encodes string from Perl's internal form into I<ENCODING> and returns |
586 | a sequence of octets. For CHECK see L</"Handling Malformed Data">. |
4411f3b6 |
587 | |
681a7c68 |
588 | For example to convert (internally UTF-8 encoded) Unicode data |
589 | to octets: |
590 | |
591 | $octets = encode("utf8", $unicode); |
592 | |
4411f3b6 |
593 | =item * |
594 | |
595 | $string = decode(ENCODING, $bytes[, CHECK]) |
596 | |
47bfe92f |
597 | Decode sequence of octets assumed to be in I<ENCODING> into Perl's |
598 | internal form and returns the resulting string. For CHECK see |
599 | L</"Handling Malformed Data">. |
600 | |
681a7c68 |
601 | For example to convert ISO 8859-1 data to UTF-8: |
602 | |
603 | $utf8 = decode("latin1", $latin1); |
604 | |
47bfe92f |
605 | =item * |
606 | |
607 | from_to($string, FROM_ENCODING, TO_ENCODING[, CHECK]) |
608 | |
2b106fbe |
609 | Convert B<in-place> the data between two encodings. How did the data |
610 | in $string originally get to be in FROM_ENCODING? Either using |
e9692b5b |
611 | encode() or through PerlIO: See L</"Encoding and IO">. For CHECK |
2b106fbe |
612 | see L</"Handling Malformed Data">. |
613 | |
614 | For example to convert ISO 8859-1 data to UTF-8: |
615 | |
616 | from_to($data, "iso-8859-1", "utf-8"); |
617 | |
618 | and to convert it back: |
619 | |
620 | from_to($data, "utf-8", "iso-8859-1"); |
4411f3b6 |
621 | |
ab97ca19 |
622 | Note that because the conversion happens in place, the data to be |
623 | converted cannot be a string constant, it must be a scalar variable. |
624 | |
4411f3b6 |
625 | =back |
626 | |
627 | =head2 Handling Malformed Data |
628 | |
629 | If CHECK is not set, C<undef> is returned. If the data is supposed to |
47bfe92f |
630 | be UTF-8, an optional lexical warning (category utf8) is given. If |
631 | CHECK is true but not a code reference, dies. |
4411f3b6 |
632 | |
47bfe92f |
633 | It would desirable to have a way to indicate that transform should use |
634 | the encodings "replacement character" - no such mechanism is defined yet. |
4411f3b6 |
635 | |
636 | It is also planned to allow I<CHECK> to be a code reference. |
637 | |
47bfe92f |
638 | This is not yet implemented as there are design issues with what its |
639 | arguments should be and how it returns its results. |
4411f3b6 |
640 | |
641 | =over 4 |
642 | |
643 | =item Scheme 1 |
644 | |
645 | Passed remaining fragment of string being processed. |
646 | Modifies it in place to remove bytes/characters it can understand |
647 | and returns a string used to represent them. |
648 | e.g. |
649 | |
650 | sub fixup { |
651 | my $ch = substr($_[0],0,1,''); |
652 | return sprintf("\x{%02X}",ord($ch); |
653 | } |
654 | |
655 | This scheme is close to how underlying C code for Encode works, but gives |
656 | the fixup routine very little context. |
657 | |
658 | =item Scheme 2 |
659 | |
47bfe92f |
660 | Passed original string, and an index into it of the problem area, and |
661 | output string so far. Appends what it will to output string and |
662 | returns new index into original string. For example: |
4411f3b6 |
663 | |
664 | sub fixup { |
665 | # my ($s,$i,$d) = @_; |
666 | my $ch = substr($_[0],$_[1],1); |
667 | $_[2] .= sprintf("\x{%02X}",ord($ch); |
668 | return $_[1]+1; |
669 | } |
670 | |
47bfe92f |
671 | This scheme gives maximal control to the fixup routine but is more |
672 | complicated to code, and may need internals of Encode to be tweaked to |
673 | keep original string intact. |
4411f3b6 |
674 | |
675 | =item Other Schemes |
676 | |
677 | Hybrids of above. |
678 | |
679 | Multiple return values rather than in-place modifications. |
680 | |
681 | Index into the string could be pos($str) allowing s/\G...//. |
682 | |
683 | =back |
684 | |
685 | =head2 UTF-8 / utf8 |
686 | |
687 | The Unicode consortium defines the UTF-8 standard as a way of encoding |
47bfe92f |
688 | the entire Unicode repertiore as sequences of octets. This encoding is |
689 | expected to become very widespread. Perl can use this form internaly |
690 | to represent strings, so conversions to and from this form are |
691 | particularly efficient (as octets in memory do not have to change, |
692 | just the meta-data that tells Perl how to treat them). |
4411f3b6 |
693 | |
694 | =over 4 |
695 | |
696 | =item * |
697 | |
698 | $bytes = encode_utf8($string); |
699 | |
47bfe92f |
700 | The characters that comprise string are encoded in Perl's superset of UTF-8 |
4411f3b6 |
701 | and the resulting octets returned as a sequence of bytes. All possible |
702 | characters have a UTF-8 representation so this function cannot fail. |
703 | |
704 | =item * |
705 | |
706 | $string = decode_utf8($bytes [,CHECK]); |
707 | |
47bfe92f |
708 | The sequence of octets represented by $bytes is decoded from UTF-8 |
709 | into a sequence of logical characters. Not all sequences of octets |
710 | form valid UTF-8 encodings, so it is possible for this call to fail. |
711 | For CHECK see L</"Handling Malformed Data">. |
4411f3b6 |
712 | |
713 | =back |
714 | |
715 | =head2 Other Encodings of Unicode |
716 | |
47bfe92f |
717 | UTF-16 is similar to UCS-2, 16 bit or 2-byte chunks. UCS-2 can only |
7a4efbb2 |
718 | represent 0..0xFFFF, while UTF-16 has a I<surrogate pair> scheme which |
47bfe92f |
719 | allows it to cover the whole Unicode range. |
4411f3b6 |
720 | |
7a4efbb2 |
721 | Surrogates are code points set aside to encode the 0x01000..0x10FFFF |
722 | range of Unicode code points in pairs of 16-bit units. The I<high |
723 | surrogates> are the range 0xD800..0xDBFF, and the I<low surrogates> |
724 | are the range 0xDC00..0xDFFFF. The surrogate encoding is |
725 | |
726 | $hi = ($uni - 0x10000) / 0x400 + 0xD800; |
727 | $lo = ($uni - 0x10000) % 0x400 + 0xDC00; |
728 | |
729 | and the decoding is |
730 | |
731 | $uni = 0x10000 + ($hi - 0xD8000) * 0x400 + ($lo - 0xDC00); |
732 | |
8040349a |
733 | Encode implements big-endian UCS-2 aliased to "iso-10646-1" as that |
47bfe92f |
734 | happens to be the name used by that representation when used with X11 |
735 | fonts. |
4411f3b6 |
736 | |
737 | UTF-32 or UCS-4 is 32-bit or 4-byte chunks. Perl's logical characters |
738 | can be considered as being in this form without encoding. An encoding |
47bfe92f |
739 | to transfer strings in this form (e.g. to write them to a file) would |
740 | need to |
4411f3b6 |
741 | |
c079d275 |
742 | pack('L*', unpack('U*', $string)); # native |
4411f3b6 |
743 | or |
c079d275 |
744 | pack('V*', unpack('U*', $string)); # little-endian |
4411f3b6 |
745 | or |
c079d275 |
746 | pack('N*', unpack('U*', $string)); # big-endian |
4411f3b6 |
747 | |
c079d275 |
748 | depending on the endianness required. |
4411f3b6 |
749 | |
51ef4e11 |
750 | No UTF-32 encodings are implemented yet. |
4411f3b6 |
751 | |
47bfe92f |
752 | Both UCS-2 and UCS-4 style encodings can have "byte order marks" by |
753 | representing the code point 0xFFFE as the very first thing in a file. |
4411f3b6 |
754 | |
51ef4e11 |
755 | =head2 Listing available encodings |
756 | |
757 | use Encode qw(encodings); |
758 | @list = encodings(); |
759 | |
760 | Returns a list of the canonical names of the available encodings. |
761 | |
762 | =head2 Defining Aliases |
763 | |
764 | use Encode qw(define_alias); |
765 | define_alias( newName => ENCODING); |
766 | |
47bfe92f |
767 | Allows newName to be used as am alias for ENCODING. ENCODING may be |
768 | either the name of an encoding or and encoding object (as above). |
51ef4e11 |
769 | |
770 | Currently I<newName> can be specified in the following ways: |
771 | |
772 | =over 4 |
773 | |
774 | =item As a simple string. |
775 | |
776 | =item As a qr// compiled regular expression, e.g.: |
777 | |
778 | define_alias( qr/^iso8859-(\d+)$/i => '"iso-8859-$1"' ); |
779 | |
47bfe92f |
780 | In this case if I<ENCODING> is not a reference it is C<eval>-ed to |
781 | allow C<$1> etc. to be subsituted. The example is one way to names as |
782 | used in X11 font names to alias the MIME names for the iso-8859-* |
783 | family. |
51ef4e11 |
784 | |
785 | =item As a code reference, e.g.: |
786 | |
787 | define_alias( sub { return /^iso8859-(\d+)$/i ? "iso-8859-$1" : undef } , ''); |
788 | |
789 | In this case C<$_> will be set to the name that is being looked up and |
47bfe92f |
790 | I<ENCODING> is passed to the sub as its first argument. The example |
791 | is another way to names as used in X11 font names to alias the MIME |
792 | names for the iso-8859-* family. |
51ef4e11 |
793 | |
794 | =back |
795 | |
796 | =head2 Defining Encodings |
797 | |
e9692b5b |
798 | use Encode qw(define_alias); |
799 | define_encoding( $object, 'canonicalName' [,alias...]); |
51ef4e11 |
800 | |
47bfe92f |
801 | Causes I<canonicalName> to be associated with I<$object>. The object |
802 | should provide the interface described in L</"IMPLEMENTATION CLASSES"> |
803 | below. If more than two arguments are provided then additional |
804 | arguments are taken as aliases for I<$object> as for C<define_alias>. |
51ef4e11 |
805 | |
4411f3b6 |
806 | =head1 Encoding and IO |
807 | |
808 | It is very common to want to do encoding transformations when |
809 | reading or writing files, network connections, pipes etc. |
47bfe92f |
810 | If Perl is configured to use the new 'perlio' IO system then |
4411f3b6 |
811 | C<Encode> provides a "layer" (See L<perliol>) which can transform |
812 | data as it is read or written. |
813 | |
8e86646e |
814 | Here is how the blind poet would modernise the encoding: |
815 | |
42234700 |
816 | use Encode; |
8e86646e |
817 | open(my $iliad,'<:encoding(iso-8859-7)','iliad.greek'); |
818 | open(my $utf8,'>:utf8','iliad.utf8'); |
819 | my @epic = <$iliad>; |
820 | print $utf8 @epic; |
821 | close($utf8); |
822 | close($illiad); |
4411f3b6 |
823 | |
824 | In addition the new IO system can also be configured to read/write |
825 | UTF-8 encoded characters (as noted above this is efficient): |
826 | |
e9692b5b |
827 | open(my $fh,'>:utf8','anything'); |
828 | print $fh "Any \x{0021} string \N{SMILEY FACE}\n"; |
4411f3b6 |
829 | |
830 | Either of the above forms of "layer" specifications can be made the default |
831 | for a lexical scope with the C<use open ...> pragma. See L<open>. |
832 | |
833 | Once a handle is open is layers can be altered using C<binmode>. |
834 | |
47bfe92f |
835 | Without any such configuration, or if Perl itself is built using |
4411f3b6 |
836 | system's own IO, then write operations assume that file handle accepts |
837 | only I<bytes> and will C<die> if a character larger than 255 is |
838 | written to the handle. When reading, each octet from the handle |
839 | becomes a byte-in-a-character. Note that this default is the same |
47bfe92f |
840 | behaviour as bytes-only languages (including Perl before v5.6) would |
841 | have, and is sufficient to handle native 8-bit encodings |
842 | e.g. iso-8859-1, EBCDIC etc. and any legacy mechanisms for handling |
843 | other encodings and binary data. |
844 | |
845 | In other cases it is the programs responsibility to transform |
846 | characters into bytes using the API above before doing writes, and to |
847 | transform the bytes read from a handle into characters before doing |
848 | "character operations" (e.g. C<lc>, C</\W+/>, ...). |
849 | |
47bfe92f |
850 | You can also use PerlIO to convert larger amounts of data you don't |
851 | want to bring into memory. For example to convert between ISO 8859-1 |
852 | (Latin 1) and UTF-8 (or UTF-EBCDIC in EBCDIC machines): |
853 | |
e9692b5b |
854 | open(F, "<:encoding(iso-8859-1)", "data.txt") or die $!; |
855 | open(G, ">:utf8", "data.utf") or die $!; |
856 | while (<F>) { print G } |
857 | |
858 | # Could also do "print G <F>" but that would pull |
859 | # the whole file into memory just to write it out again. |
860 | |
861 | More examples: |
47bfe92f |
862 | |
e9692b5b |
863 | open(my $f, "<:encoding(cp1252)") |
864 | open(my $g, ">:encoding(iso-8859-2)") |
865 | open(my $h, ">:encoding(latin9)") # iso-8859-15 |
47bfe92f |
866 | |
867 | See L<PerlIO> for more information. |
4411f3b6 |
868 | |
1768d7eb |
869 | See also L<encoding> for how to change the default encoding of the |
d521382b |
870 | data in your script. |
1768d7eb |
871 | |
4411f3b6 |
872 | =head1 Encoding How to ... |
873 | |
874 | To do: |
875 | |
876 | =over 4 |
877 | |
878 | =item * IO with mixed content (faking iso-2020-*) |
879 | |
880 | =item * MIME's Content-Length: |
881 | |
882 | =item * UTF-8 strings in binary data. |
883 | |
47bfe92f |
884 | =item * Perl/Encode wrappers on non-Unicode XS modules. |
4411f3b6 |
885 | |
886 | =back |
887 | |
888 | =head1 Messing with Perl's Internals |
889 | |
47bfe92f |
890 | The following API uses parts of Perl's internals in the current |
891 | implementation. As such they are efficient, but may change. |
4411f3b6 |
892 | |
893 | =over 4 |
894 | |
4411f3b6 |
895 | =item * is_utf8(STRING [, CHECK]) |
896 | |
897 | [INTERNAL] Test whether the UTF-8 flag is turned on in the STRING. |
47bfe92f |
898 | If CHECK is true, also checks the data in STRING for being well-formed |
899 | UTF-8. Returns true if successful, false otherwise. |
4411f3b6 |
900 | |
901 | =item * valid_utf8(STRING) |
902 | |
47bfe92f |
903 | [INTERNAL] Test whether STRING is in a consistent state. Will return |
904 | true if string is held as bytes, or is well-formed UTF-8 and has the |
905 | UTF-8 flag on. Main reason for this routine is to allow Perl's |
906 | testsuite to check that operations have left strings in a consistent |
907 | state. |
4411f3b6 |
908 | |
909 | =item * |
910 | |
911 | _utf8_on(STRING) |
912 | |
913 | [INTERNAL] Turn on the UTF-8 flag in STRING. The data in STRING is |
914 | B<not> checked for being well-formed UTF-8. Do not use unless you |
915 | B<know> that the STRING is well-formed UTF-8. Returns the previous |
916 | state of the UTF-8 flag (so please don't test the return value as |
917 | I<not> success or failure), or C<undef> if STRING is not a string. |
918 | |
919 | =item * |
920 | |
921 | _utf8_off(STRING) |
922 | |
923 | [INTERNAL] Turn off the UTF-8 flag in STRING. Do not use frivolously. |
924 | Returns the previous state of the UTF-8 flag (so please don't test the |
925 | return value as I<not> success or failure), or C<undef> if STRING is |
926 | not a string. |
927 | |
928 | =back |
929 | |
4edaa979 |
930 | =head1 IMPLEMENTATION CLASSES |
931 | |
932 | As mentioned above encodings are (in the current implementation at least) |
933 | defined by objects. The mapping of encoding name to object is via the |
51ef4e11 |
934 | C<%encodings> hash. |
4edaa979 |
935 | |
936 | The values of the hash can currently be either strings or objects. |
937 | The string form may go away in the future. The string form occurs |
938 | when C<encodings()> has scanned C<@INC> for loadable encodings but has |
939 | not actually loaded the encoding in question. This is because the |
47bfe92f |
940 | current "loading" process is all Perl and a bit slow. |
4edaa979 |
941 | |
47bfe92f |
942 | Once an encoding is loaded then value of the hash is object which |
943 | implements the encoding. The object should provide the following |
944 | interface: |
4edaa979 |
945 | |
946 | =over 4 |
947 | |
948 | =item -E<gt>name |
949 | |
950 | Should return the string representing the canonical name of the encoding. |
951 | |
952 | =item -E<gt>new_sequence |
953 | |
47bfe92f |
954 | This is a placeholder for encodings with state. It should return an |
955 | object which implements this interface, all current implementations |
956 | return the original object. |
4edaa979 |
957 | |
958 | =item -E<gt>encode($string,$check) |
959 | |
47bfe92f |
960 | Should return the octet sequence representing I<$string>. If I<$check> |
961 | is true it should modify I<$string> in place to remove the converted |
962 | part (i.e. the whole string unless there is an error). If an error |
963 | occurs it should return the octet sequence for the fragment of string |
964 | that has been converted, and modify $string in-place to remove the |
965 | converted part leaving it starting with the problem fragment. |
4edaa979 |
966 | |
47bfe92f |
967 | If check is is false then C<encode> should make a "best effort" to |
968 | convert the string - for example by using a replacement character. |
4edaa979 |
969 | |
970 | =item -E<gt>decode($octets,$check) |
971 | |
47bfe92f |
972 | Should return the string that I<$octets> represents. If I<$check> is |
973 | true it should modify I<$octets> in place to remove the converted part |
974 | (i.e. the whole sequence unless there is an error). If an error |
975 | occurs it should return the fragment of string that has been |
976 | converted, and modify $octets in-place to remove the converted part |
4edaa979 |
977 | leaving it starting with the problem fragment. |
978 | |
47bfe92f |
979 | If check is is false then C<decode> should make a "best effort" to |
980 | convert the string - for example by using Unicode's "\x{FFFD}" as a |
981 | replacement character. |
4edaa979 |
982 | |
983 | =back |
984 | |
47bfe92f |
985 | It should be noted that the check behaviour is different from the |
986 | outer public API. The logic is that the "unchecked" case is useful |
987 | when encoding is part of a stream which may be reporting errors |
988 | (e.g. STDERR). In such cases it is desirable to get everything |
989 | through somehow without causing additional errors which obscure the |
990 | original one. Also the encoding is best placed to know what the |
991 | correct replacement character is, so if that is the desired behaviour |
992 | then letting low level code do it is the most efficient. |
993 | |
994 | In contrast if check is true, the scheme above allows the encoding to |
995 | do as much as it can and tell layer above how much that was. What is |
996 | lacking at present is a mechanism to report what went wrong. The most |
997 | likely interface will be an additional method call to the object, or |
998 | perhaps (to avoid forcing per-stream objects on otherwise stateless |
999 | encodings) and additional parameter. |
1000 | |
1001 | It is also highly desirable that encoding classes inherit from |
1002 | C<Encode::Encoding> as a base class. This allows that class to define |
1003 | additional behaviour for all encoding objects. For example built in |
1004 | Unicode, UCS-2 and UTF-8 classes use : |
51ef4e11 |
1005 | |
1006 | package Encode::MyEncoding; |
1007 | use base qw(Encode::Encoding); |
1008 | |
1009 | __PACKAGE__->Define(qw(myCanonical myAlias)); |
1010 | |
47bfe92f |
1011 | To create an object with bless {Name => ...},$class, and call |
1012 | define_encoding. They inherit their C<name> method from |
1013 | C<Encode::Encoding>. |
4edaa979 |
1014 | |
1015 | =head2 Compiled Encodings |
1016 | |
47bfe92f |
1017 | F<Encode.xs> provides a class C<Encode::XS> which provides the |
1018 | interface described above. It calls a generic octet-sequence to |
1019 | octet-sequence "engine" that is driven by tables (defined in |
1020 | F<encengine.c>). The same engine is used for both encode and |
1021 | decode. C<Encode:XS>'s C<encode> forces Perl's characters to their |
1022 | UTF-8 form and then treats them as just another multibyte |
1023 | encoding. C<Encode:XS>'s C<decode> transforms the sequence and then |
1024 | turns the UTF-8-ness flag as that is the form that the tables are |
1025 | defined to produce. For details of the engine see the comments in |
1026 | F<encengine.c>. |
1027 | |
1028 | The tables are produced by the Perl script F<compile> (the name needs |
1029 | to change so we can eventually install it somewhere). F<compile> can |
1030 | currently read two formats: |
4edaa979 |
1031 | |
1032 | =over 4 |
1033 | |
1034 | =item *.enc |
1035 | |
47bfe92f |
1036 | This is a coined format used by Tcl. It is documented in |
1037 | Encode/EncodeFormat.pod. |
4edaa979 |
1038 | |
1039 | =item *.ucm |
1040 | |
1041 | This is the semi-standard format used by IBM's ICU package. |
1042 | |
1043 | =back |
1044 | |
1045 | F<compile> can write the following forms: |
1046 | |
1047 | =over 4 |
1048 | |
1049 | =item *.ucm |
1050 | |
1051 | See above - the F<Encode/*.ucm> files provided with the distribution have |
1052 | been created from the original Tcl .enc files using this approach. |
1053 | |
1054 | =item *.c |
1055 | |
1056 | Produces tables as C data structures - this is used to build in encodings |
1057 | into F<Encode.so>/F<Encode.dll>. |
1058 | |
1059 | =item *.xs |
1060 | |
47bfe92f |
1061 | In theory this allows encodings to be stand-alone loadable Perl |
1062 | extensions. The process has not yet been tested. The plan is to use |
1063 | this approach for large East Asian encodings. |
4edaa979 |
1064 | |
1065 | =back |
1066 | |
47bfe92f |
1067 | The set of encodings built-in to F<Encode.so>/F<Encode.dll> is |
1068 | determined by F<Makefile.PL>. The current set is as follows: |
4edaa979 |
1069 | |
1070 | =over 4 |
1071 | |
1072 | =item ascii and iso-8859-* |
1073 | |
1074 | That is all the common 8-bit "western" encodings. |
1075 | |
1076 | =item IBM-1047 and two other variants of EBCDIC. |
1077 | |
47bfe92f |
1078 | These are the same variants that are supported by EBCDIC Perl as |
1079 | "native" encodings. They are included to prove "reversibility" of |
1080 | some constructs in EBCDIC Perl. |
4edaa979 |
1081 | |
1082 | =item symbol and dingbats as used by Tk on X11. |
1083 | |
47bfe92f |
1084 | (The reason Encode got started was to support Perl/Tk.) |
4edaa979 |
1085 | |
1086 | =back |
1087 | |
47bfe92f |
1088 | That set is rather ad hoc and has been driven by the needs of the |
1089 | tests rather than the needs of typical applications. It is likely |
1090 | to be rationalized. |
4edaa979 |
1091 | |
4411f3b6 |
1092 | =head1 SEE ALSO |
1093 | |
1768d7eb |
1094 | L<perlunicode>, L<perlebcdic>, L<perlfunc/open>, L<PerlIO>, L<encoding> |
4411f3b6 |
1095 | |
1096 | =cut |
1097 | |