Commit | Line | Data |
2c674647 |
1 | package Encode; |
51ef4e11 |
2 | use strict; |
2c674647 |
3 | |
b8a524e9 |
4 | our $VERSION = '0.02'; |
2c674647 |
5 | |
6 | require DynaLoader; |
7 | require Exporter; |
8 | |
51ef4e11 |
9 | our @ISA = qw(Exporter DynaLoader); |
2c674647 |
10 | |
4411f3b6 |
11 | # Public, encouraged API is exported by default |
51ef4e11 |
12 | our @EXPORT = qw ( |
4411f3b6 |
13 | encode |
14 | decode |
15 | encode_utf8 |
16 | decode_utf8 |
17 | find_encoding |
51ef4e11 |
18 | encodings |
4411f3b6 |
19 | ); |
20 | |
51ef4e11 |
21 | our @EXPORT_OK = |
2c674647 |
22 | qw( |
51ef4e11 |
23 | define_encoding |
24 | define_alias |
2c674647 |
25 | from_to |
26 | is_utf8 |
4411f3b6 |
27 | is_8bit |
28 | is_16bit |
a12c0f56 |
29 | utf8_upgrade |
30 | utf8_downgrade |
4411f3b6 |
31 | _utf8_on |
32 | _utf8_off |
2c674647 |
33 | ); |
34 | |
35 | bootstrap Encode (); |
36 | |
4411f3b6 |
37 | # Documentation moved after __END__ for speed - NI-S |
2c674647 |
38 | |
bf230f3d |
39 | use Carp; |
40 | |
51ef4e11 |
41 | # Make a %encoding package variable to allow a certain amount of cheating |
42 | our %encoding; |
43 | my @alias; # ordered matching list |
44 | my %alias; # cached known aliases |
f7ac3676 |
45 | |
6d6a7c8d |
46 | # 0 1 2 3 4 5 6 7 8 9 10 |
47 | our @latin2iso_num = ( 0, 1, 2, 3, 4, 9, 10, 13, 14, 15, 16 ); |
48 | |
f7ac3676 |
49 | our %winlatin2cp = ( |
50 | 'Latin1' => 1252, |
51 | 'Latin2' => 1250, |
52 | 'Cyrillic' => 1251, |
53 | 'Baltic' => 1257, |
54 | 'Greek' => 1253, |
55 | 'Turkish' => 1254, |
56 | 'Hebrew' => 1255, |
57 | 'Arabic' => 1256, |
58 | 'Baltic' => 1257, |
59 | 'Vietnamese' => 1258, |
60 | ); |
5345d506 |
61 | |
656753f8 |
62 | sub encodings |
63 | { |
64 | my ($class) = @_; |
51ef4e11 |
65 | return keys %encoding; |
66 | } |
67 | |
68 | sub findAlias |
69 | { |
70 | my $class = shift; |
71 | local $_ = shift; |
1e616cf5 |
72 | # print "# findAlias $_\n"; |
51ef4e11 |
73 | unless (exists $alias{$_}) |
656753f8 |
74 | { |
51ef4e11 |
75 | for (my $i=0; $i < @alias; $i += 2) |
656753f8 |
76 | { |
51ef4e11 |
77 | my $alias = $alias[$i]; |
78 | my $val = $alias[$i+1]; |
79 | my $new; |
80 | if (ref($alias) eq 'Regexp' && $_ =~ $alias) |
5345d506 |
81 | { |
51ef4e11 |
82 | $new = eval $val; |
83 | } |
84 | elsif (ref($alias) eq 'CODE') |
85 | { |
86 | $new = &{$alias}($val) |
87 | } |
5ad8ef52 |
88 | elsif (lc($_) eq lc($alias)) |
51ef4e11 |
89 | { |
90 | $new = $val; |
91 | } |
92 | if (defined($new)) |
93 | { |
94 | next if $new eq $_; # avoid (direct) recursion on bugs |
95 | my $enc = (ref($new)) ? $new : find_encoding($new); |
96 | if ($enc) |
5345d506 |
97 | { |
51ef4e11 |
98 | $alias{$_} = $enc; |
99 | last; |
5345d506 |
100 | } |
101 | } |
656753f8 |
102 | } |
5345d506 |
103 | } |
51ef4e11 |
104 | return $alias{$_}; |
5345d506 |
105 | } |
106 | |
51ef4e11 |
107 | sub define_alias |
5345d506 |
108 | { |
51ef4e11 |
109 | while (@_) |
5345d506 |
110 | { |
51ef4e11 |
111 | my ($alias,$name) = splice(@_,0,2); |
112 | push(@alias, $alias => $name); |
656753f8 |
113 | } |
51ef4e11 |
114 | } |
115 | |
016cb72c |
116 | # Allow variants of iso-8859-1 etc. |
d6089a2a |
117 | define_alias( qr/^iso[-_]?(\d+)[-_](\d+)$/i => '"iso-$1-$2"' ); |
016cb72c |
118 | |
7faf300d |
119 | # At least HP-UX has these. |
120 | define_alias( qr/^iso8859(\d+)$/i => '"iso-8859-$1"' ); |
121 | |
f7ac3676 |
122 | # More HP stuff. |
123 | define_alias( qr/^(?:hp-)?(arabic|greek|hebrew|kana|roman|thai|turkish)8$/i => '"${1}8"' ); |
124 | |
8a361256 |
125 | # The Official name of ASCII. |
126 | define_alias( qr/^ANSI[-_]?X3\.4[-_]?1968$/i => '"ascii"' ); |
127 | |
58d53262 |
128 | # This is a font issue, not an encoding issue. |
129 | # (The currency symbol of the Latin 1 upper half |
130 | # has been redefined as the euro symbol.) |
131 | define_alias( qr/^(.+)\@euro$/i => '"$1"' ); |
132 | |
016cb72c |
133 | # Allow latin-1 style names as well |
7faf300d |
134 | define_alias( qr/^(?:iso[-_]?)?latin[-_]?(\d+)$/i => '"iso-8859-$latin2iso_num[$1]"' ); |
016cb72c |
135 | |
f7ac3676 |
136 | # Allow winlatin1 style names as well |
cf91068f |
137 | define_alias( qr/^win(latin[12]|cyrillic|baltic|greek|turkish|hebrew|arabic|baltic|vietnamese)$/i => '"cp$winlatin2cp{\u$1}"' ); |
f7ac3676 |
138 | |
016cb72c |
139 | # Common names for non-latin prefered MIME names |
140 | define_alias( 'ascii' => 'US-ascii', |
141 | 'cyrillic' => 'iso-8859-5', |
142 | 'arabic' => 'iso-8859-6', |
143 | 'greek' => 'iso-8859-7', |
f7ac3676 |
144 | 'hebrew' => 'iso-8859-8', |
145 | 'thai' => 'iso-8859-11', |
146 | 'tis620' => 'iso-8859-11', |
147 | ); |
016cb72c |
148 | |
7faf300d |
149 | # At least AIX has IBM-NNN (surprisingly...) instead of cpNNN. |
150 | define_alias( qr/^ibm[-_]?(\d\d\d\d?)$/i => '"cp$1"'); |
151 | |
58d53262 |
152 | # Standardize on the dashed versions. |
153 | define_alias( qr/^utf8$/i => 'utf-8' ); |
7faf300d |
154 | define_alias( qr/^koi8r$/i => 'koi8-r' ); |
f7ac3676 |
155 | define_alias( qr/^koi8u$/i => 'koi8-u' ); |
156 | |
157 | # TODO: HP-UX '8' encodings arabic8 greek8 hebrew8 kana8 thai8 turkish8 |
158 | # TODO: HP-UX '15' encodings japanese15 korean15 roi15 |
159 | # TODO: Cyrillic encoding ISO-IR-111 (useful?) |
160 | # TODO: Chinese encodings GB18030 GBK Big5-HSKCS EUC-TW |
161 | # TODO: Armenian encoding ARMSCII-8 |
162 | # TODO: Hebrew encoding ISO-8859-8-1 |
163 | # TODO: Thai encoding TCVN |
164 | # TODO: Korean encoding Johab |
165 | # TODO: Vietnamese encodings VISCII VPS |
166 | # TODO: Japanese encoding JIS (not the same as SJIS) |
167 | # TODO: Mac Asian+African encodings: Arabic Armenian Bengali Burmese |
168 | # ChineseSimp ChineseTrad Devanagari Ethiopic ExtArabic |
169 | # Farsi Georgian Gujarati Gurmukhi Hebrew Japanese |
170 | # Kannada Khmer Korean Laotian Malayalam Mongolian |
171 | # Oriya Sinhalese Symbol Tamil Telugu Tibetan Vietnamese |
172 | # TODO: what is the Japanese 'UJIS' encoding seen in some Linuxes? |
58d53262 |
173 | |
016cb72c |
174 | # Map white space and _ to '-' |
175 | define_alias( qr/^(\S+)[\s_]+(.*)$/i => '"$1-$2"' ); |
176 | |
51ef4e11 |
177 | sub define_encoding |
178 | { |
179 | my $obj = shift; |
180 | my $name = shift; |
181 | $encoding{$name} = $obj; |
182 | my $lc = lc($name); |
183 | define_alias($lc => $obj) unless $lc eq $name; |
184 | while (@_) |
656753f8 |
185 | { |
51ef4e11 |
186 | my $alias = shift; |
187 | define_alias($alias,$obj); |
656753f8 |
188 | } |
51ef4e11 |
189 | return $obj; |
656753f8 |
190 | } |
191 | |
656753f8 |
192 | sub getEncoding |
193 | { |
194 | my ($class,$name) = @_; |
5345d506 |
195 | my $enc; |
0f43fc90 |
196 | if (ref($name) && $name->can('new_sequence')) |
197 | { |
198 | return $name; |
199 | } |
1e616cf5 |
200 | my $lc = lc $name; |
51ef4e11 |
201 | if (exists $encoding{$name}) |
656753f8 |
202 | { |
51ef4e11 |
203 | return $encoding{$name}; |
204 | } |
1e616cf5 |
205 | if (exists $encoding{$lc}) |
51ef4e11 |
206 | { |
1e616cf5 |
207 | return $encoding{$lc}; |
656753f8 |
208 | } |
1e616cf5 |
209 | |
210 | my $oc = $class->findAlias($name); |
211 | return $oc if defined $oc; |
212 | return $class->findAlias($lc) if $lc ne $name; |
213 | |
214 | return; |
656753f8 |
215 | } |
216 | |
4411f3b6 |
217 | sub find_encoding |
218 | { |
219 | my ($name) = @_; |
220 | return __PACKAGE__->getEncoding($name); |
221 | } |
222 | |
223 | sub encode |
224 | { |
225 | my ($name,$string,$check) = @_; |
226 | my $enc = find_encoding($name); |
227 | croak("Unknown encoding '$name'") unless defined $enc; |
50d26985 |
228 | my $octets = $enc->encode($string,$check); |
4411f3b6 |
229 | return undef if ($check && length($string)); |
230 | return $octets; |
231 | } |
232 | |
233 | sub decode |
234 | { |
235 | my ($name,$octets,$check) = @_; |
236 | my $enc = find_encoding($name); |
237 | croak("Unknown encoding '$name'") unless defined $enc; |
50d26985 |
238 | my $string = $enc->decode($octets,$check); |
96d6357c |
239 | $_[1] = $octets if $check; |
4411f3b6 |
240 | return $string; |
241 | } |
242 | |
243 | sub from_to |
244 | { |
245 | my ($string,$from,$to,$check) = @_; |
246 | my $f = find_encoding($from); |
247 | croak("Unknown encoding '$from'") unless defined $f; |
248 | my $t = find_encoding($to); |
249 | croak("Unknown encoding '$to'") unless defined $t; |
50d26985 |
250 | my $uni = $f->decode($string,$check); |
4411f3b6 |
251 | return undef if ($check && length($string)); |
50d26985 |
252 | $string = $t->encode($uni,$check); |
4411f3b6 |
253 | return undef if ($check && length($uni)); |
254 | return length($_[0] = $string); |
255 | } |
256 | |
257 | sub encode_utf8 |
258 | { |
259 | my ($str) = @_; |
1b026014 |
260 | utf8::encode($str); |
4411f3b6 |
261 | return $str; |
262 | } |
263 | |
264 | sub decode_utf8 |
265 | { |
266 | my ($str) = @_; |
1b026014 |
267 | return undef unless utf8::decode($str); |
4411f3b6 |
268 | return $str; |
269 | } |
270 | |
50d26985 |
271 | package Encode::Encoding; |
272 | # Base class for classes which implement encodings |
4edaa979 |
273 | |
51ef4e11 |
274 | sub Define |
275 | { |
276 | my $obj = shift; |
277 | my $canonical = shift; |
278 | $obj = bless { Name => $canonical },$obj unless ref $obj; |
279 | # warn "$canonical => $obj\n"; |
280 | Encode::define_encoding($obj, $canonical, @_); |
281 | } |
282 | |
283 | sub name { shift->{'Name'} } |
284 | |
50d26985 |
285 | # Temporary legacy methods |
4edaa979 |
286 | sub toUnicode { shift->decode(@_) } |
287 | sub fromUnicode { shift->encode(@_) } |
288 | |
289 | sub new_sequence { return $_[0] } |
50d26985 |
290 | |
291 | package Encode::XS; |
292 | use base 'Encode::Encoding'; |
293 | |
5ad8ef52 |
294 | package Encode::Internal; |
50d26985 |
295 | use base 'Encode::Encoding'; |
656753f8 |
296 | |
9b37254d |
297 | # Dummy package that provides the encode interface but leaves data |
1b026014 |
298 | # as UTF-X encoded. It is here so that from_to() works. |
656753f8 |
299 | |
5ad8ef52 |
300 | __PACKAGE__->Define('Internal'); |
301 | |
302 | Encode::define_alias( 'Unicode' => 'Internal' ) if ord('A') == 65; |
656753f8 |
303 | |
50d26985 |
304 | sub decode |
a12c0f56 |
305 | { |
306 | my ($obj,$str,$chk) = @_; |
1b026014 |
307 | utf8::upgrade($str); |
a12c0f56 |
308 | $_[1] = '' if $chk; |
309 | return $str; |
310 | } |
656753f8 |
311 | |
50d26985 |
312 | *encode = \&decode; |
656753f8 |
313 | |
5ad8ef52 |
314 | package Encoding::Unicode; |
315 | use base 'Encode::Encoding'; |
316 | |
317 | __PACKAGE__->Define('Unicode') unless ord('A') == 65; |
318 | |
319 | sub decode |
320 | { |
321 | my ($obj,$str,$chk) = @_; |
322 | my $res = ''; |
323 | for (my $i = 0; $i < length($str); $i++) |
324 | { |
325 | $res .= chr(utf8::unicode_to_native(ord(substr($str,$i,1)))); |
326 | } |
327 | $_[1] = '' if $chk; |
328 | return $res; |
329 | } |
330 | |
331 | sub encode |
332 | { |
333 | my ($obj,$str,$chk) = @_; |
334 | my $res = ''; |
335 | for (my $i = 0; $i < length($str); $i++) |
336 | { |
337 | $res .= chr(utf8::native_to_unicode(ord(substr($str,$i,1)))); |
338 | } |
339 | $_[1] = '' if $chk; |
340 | return $res; |
341 | } |
342 | |
343 | |
4411f3b6 |
344 | package Encode::utf8; |
50d26985 |
345 | use base 'Encode::Encoding'; |
4411f3b6 |
346 | # package to allow long-hand |
347 | # $octets = encode( utf8 => $string ); |
348 | # |
349 | |
51ef4e11 |
350 | __PACKAGE__->Define(qw(UTF-8 utf8)); |
4411f3b6 |
351 | |
50d26985 |
352 | sub decode |
4411f3b6 |
353 | { |
354 | my ($obj,$octets,$chk) = @_; |
2a936312 |
355 | my $str = Encode::decode_utf8($octets); |
4411f3b6 |
356 | if (defined $str) |
357 | { |
358 | $_[1] = '' if $chk; |
359 | return $str; |
360 | } |
361 | return undef; |
362 | } |
363 | |
50d26985 |
364 | sub encode |
4411f3b6 |
365 | { |
366 | my ($obj,$string,$chk) = @_; |
2a936312 |
367 | my $octets = Encode::encode_utf8($string); |
4411f3b6 |
368 | $_[1] = '' if $chk; |
369 | return $octets; |
4411f3b6 |
370 | } |
371 | |
9b37254d |
372 | package Encode::iso10646_1; |
50d26985 |
373 | use base 'Encode::Encoding'; |
51ef4e11 |
374 | # Encoding is 16-bit network order Unicode (no surogates) |
9b37254d |
375 | # Used for X font encodings |
87714904 |
376 | |
8040349a |
377 | __PACKAGE__->Define(qw(UCS-2 iso-10646-1)); |
87714904 |
378 | |
50d26985 |
379 | sub decode |
87714904 |
380 | { |
381 | my ($obj,$str,$chk) = @_; |
382 | my $uni = ''; |
383 | while (length($str)) |
384 | { |
5dcbab34 |
385 | my $code = unpack('n',substr($str,0,2,'')) & 0xffff; |
87714904 |
386 | $uni .= chr($code); |
387 | } |
388 | $_[1] = $str if $chk; |
8040349a |
389 | utf8::upgrade($uni); |
87714904 |
390 | return $uni; |
391 | } |
392 | |
50d26985 |
393 | sub encode |
87714904 |
394 | { |
395 | my ($obj,$uni,$chk) = @_; |
396 | my $str = ''; |
397 | while (length($uni)) |
398 | { |
399 | my $ch = substr($uni,0,1,''); |
400 | my $x = ord($ch); |
401 | unless ($x < 32768) |
402 | { |
403 | last if ($chk); |
404 | $x = 0; |
405 | } |
5dcbab34 |
406 | $str .= pack('n',$x); |
656753f8 |
407 | } |
bf230f3d |
408 | $_[1] = $uni if $chk; |
656753f8 |
409 | return $str; |
410 | } |
411 | |
79019f4f |
412 | package Encode::ucs_2le; |
413 | use base 'Encode::Encoding'; |
414 | |
415 | __PACKAGE__->Define(qw(UCS-2le UCS-2LE ucs-2le)); |
416 | |
417 | sub decode |
418 | { |
419 | my ($obj,$str,$chk) = @_; |
420 | my $uni = ''; |
421 | while (length($str)) |
422 | { |
423 | my $code = unpack('v',substr($str,0,2,'')) & 0xffff; |
424 | $uni .= chr($code); |
425 | } |
426 | $_[1] = $str if $chk; |
427 | utf8::upgrade($uni); |
428 | return $uni; |
429 | } |
430 | |
431 | sub encode |
432 | { |
433 | my ($obj,$uni,$chk) = @_; |
434 | my $str = ''; |
435 | while (length($uni)) |
436 | { |
437 | my $ch = substr($uni,0,1,''); |
438 | my $x = ord($ch); |
439 | unless ($x < 32768) |
440 | { |
441 | last if ($chk); |
442 | $x = 0; |
443 | } |
444 | $str .= pack('v',$x); |
445 | } |
446 | $_[1] = $uni if $chk; |
447 | return $str; |
448 | } |
449 | |
4411f3b6 |
450 | # switch back to Encode package in case we ever add AutoLoader |
451 | package Encode; |
452 | |
656753f8 |
453 | 1; |
454 | |
2a936312 |
455 | __END__ |
456 | |
4411f3b6 |
457 | =head1 NAME |
458 | |
459 | Encode - character encodings |
460 | |
461 | =head1 SYNOPSIS |
462 | |
463 | use Encode; |
464 | |
465 | =head1 DESCRIPTION |
466 | |
47bfe92f |
467 | The C<Encode> module provides the interfaces between Perl's strings |
468 | and the rest of the system. Perl strings are sequences of B<characters>. |
4411f3b6 |
469 | |
470 | The repertoire of characters that Perl can represent is at least that |
47bfe92f |
471 | defined by the Unicode Consortium. On most platforms the ordinal |
472 | values of the characters (as returned by C<ord(ch)>) is the "Unicode |
473 | codepoint" for the character (the exceptions are those platforms where |
474 | the legacy encoding is some variant of EBCDIC rather than a super-set |
475 | of ASCII - see L<perlebcdic>). |
4411f3b6 |
476 | |
477 | Traditionaly computer data has been moved around in 8-bit chunks |
478 | often called "bytes". These chunks are also known as "octets" in |
479 | networking standards. Perl is widely used to manipulate data of |
480 | many types - not only strings of characters representing human or |
481 | computer languages but also "binary" data being the machines representation |
482 | of numbers, pixels in an image - or just about anything. |
483 | |
47bfe92f |
484 | When Perl is processing "binary data" the programmer wants Perl to process |
485 | "sequences of bytes". This is not a problem for Perl - as a byte has 256 |
486 | possible values it easily fits in Perl's much larger "logical character". |
4411f3b6 |
487 | |
488 | =head2 TERMINOLOGY |
489 | |
4ac9195f |
490 | =over 4 |
4411f3b6 |
491 | |
492 | =item * |
493 | |
494 | I<character>: a character in the range 0..(2**32-1) (or more). |
47bfe92f |
495 | (What Perl's strings are made of.) |
4411f3b6 |
496 | |
497 | =item * |
498 | |
499 | I<byte>: a character in the range 0..255 |
47bfe92f |
500 | (A special case of a Perl character.) |
4411f3b6 |
501 | |
502 | =item * |
503 | |
504 | I<octet>: 8 bits of data, with ordinal values 0..255 |
47bfe92f |
505 | (Term for bytes passed to or from a non-Perl context, e.g. disk file.) |
4411f3b6 |
506 | |
507 | =back |
508 | |
509 | The marker [INTERNAL] marks Internal Implementation Details, in |
510 | general meant only for those who think they know what they are doing, |
511 | and such details may change in future releases. |
512 | |
513 | =head1 ENCODINGS |
514 | |
515 | =head2 Characteristics of an Encoding |
516 | |
517 | An encoding has a "repertoire" of characters that it can represent, |
518 | and for each representable character there is at least one sequence of |
519 | octets that represents it. |
520 | |
521 | =head2 Types of Encodings |
522 | |
523 | Encodings can be divided into the following types: |
524 | |
525 | =over 4 |
526 | |
527 | =item * Fixed length 8-bit (or less) encodings. |
528 | |
529 | Each character is a single octet so may have a repertoire of up to |
530 | 256 characters. ASCII and iso-8859-* are typical examples. |
531 | |
532 | =item * Fixed length 16-bit encodings |
533 | |
534 | Each character is two octets so may have a repertoire of up to |
47bfe92f |
535 | 65 536 characters. Unicode's UCS-2 is an example. Also used for |
4411f3b6 |
536 | encodings for East Asian languages. |
537 | |
538 | =item * Fixed length 32-bit encodings. |
539 | |
540 | Not really very "encoded" encodings. The Unicode code points |
541 | are just represented as 4-octet integers. None the less because |
542 | different architectures use different representations of integers |
543 | (so called "endian") there at least two disctinct encodings. |
544 | |
545 | =item * Multi-byte encodings |
546 | |
547 | The number of octets needed to represent a character varies. |
548 | UTF-8 is a particularly complex but regular case of a multi-byte |
549 | encoding. Several East Asian countries use a multi-byte encoding |
550 | where 1-octet is used to cover western roman characters and Asian |
551 | characters get 2-octets. |
552 | (UTF-16 is strictly a multi-byte encoding taking either 2 or 4 octets |
553 | to represent a Unicode code point.) |
554 | |
555 | =item * "Escape" encodings. |
556 | |
557 | These encodings embed "escape sequences" into the octet sequence |
558 | which describe how the following octets are to be interpreted. |
559 | The iso-2022-* family is typical. Following the escape sequence |
560 | octets are encoded by an "embedded" encoding (which will be one |
561 | of the above types) until another escape sequence switches to |
562 | a different "embedded" encoding. |
563 | |
564 | These schemes are very flexible and can handle mixed languages but are |
47bfe92f |
565 | very complex to process (and have state). No escape encodings are |
566 | implemented for Perl yet. |
4411f3b6 |
567 | |
568 | =back |
569 | |
570 | =head2 Specifying Encodings |
571 | |
572 | Encodings can be specified to the API described below in two ways: |
573 | |
574 | =over 4 |
575 | |
576 | =item 1. By name |
577 | |
47bfe92f |
578 | Encoding names are strings with characters taken from a restricted |
579 | repertoire. See L</"Encoding Names">. |
4411f3b6 |
580 | |
581 | =item 2. As an object |
582 | |
583 | Encoding objects are returned by C<find_encoding($name)>. |
584 | |
585 | =back |
586 | |
587 | =head2 Encoding Names |
588 | |
589 | Encoding names are case insensitive. White space in names is ignored. |
47bfe92f |
590 | In addition an encoding may have aliases. Each encoding has one |
591 | "canonical" name. The "canonical" name is chosen from the names of |
592 | the encoding by picking the first in the following sequence: |
4411f3b6 |
593 | |
594 | =over 4 |
595 | |
596 | =item * The MIME name as defined in IETF RFC-XXXX. |
597 | |
598 | =item * The name in the IANA registry. |
599 | |
d1be9408 |
600 | =item * The name used by the organization that defined it. |
4411f3b6 |
601 | |
602 | =back |
603 | |
604 | Because of all the alias issues, and because in the general case |
605 | encodings have state C<Encode> uses the encoding object internally |
606 | once an operation is in progress. |
607 | |
4411f3b6 |
608 | =head1 PERL ENCODING API |
609 | |
610 | =head2 Generic Encoding Interface |
611 | |
612 | =over 4 |
613 | |
614 | =item * |
615 | |
616 | $bytes = encode(ENCODING, $string[, CHECK]) |
617 | |
47bfe92f |
618 | Encodes string from Perl's internal form into I<ENCODING> and returns |
619 | a sequence of octets. For CHECK see L</"Handling Malformed Data">. |
4411f3b6 |
620 | |
621 | =item * |
622 | |
623 | $string = decode(ENCODING, $bytes[, CHECK]) |
624 | |
47bfe92f |
625 | Decode sequence of octets assumed to be in I<ENCODING> into Perl's |
626 | internal form and returns the resulting string. For CHECK see |
627 | L</"Handling Malformed Data">. |
628 | |
629 | =item * |
630 | |
631 | from_to($string, FROM_ENCODING, TO_ENCODING[, CHECK]) |
632 | |
2b106fbe |
633 | Convert B<in-place> the data between two encodings. How did the data |
634 | in $string originally get to be in FROM_ENCODING? Either using |
e9692b5b |
635 | encode() or through PerlIO: See L</"Encoding and IO">. For CHECK |
2b106fbe |
636 | see L</"Handling Malformed Data">. |
637 | |
638 | For example to convert ISO 8859-1 data to UTF-8: |
639 | |
640 | from_to($data, "iso-8859-1", "utf-8"); |
641 | |
642 | and to convert it back: |
643 | |
644 | from_to($data, "utf-8", "iso-8859-1"); |
4411f3b6 |
645 | |
ab97ca19 |
646 | Note that because the conversion happens in place, the data to be |
647 | converted cannot be a string constant, it must be a scalar variable. |
648 | |
4411f3b6 |
649 | =back |
650 | |
651 | =head2 Handling Malformed Data |
652 | |
653 | If CHECK is not set, C<undef> is returned. If the data is supposed to |
47bfe92f |
654 | be UTF-8, an optional lexical warning (category utf8) is given. If |
655 | CHECK is true but not a code reference, dies. |
4411f3b6 |
656 | |
47bfe92f |
657 | It would desirable to have a way to indicate that transform should use |
658 | the encodings "replacement character" - no such mechanism is defined yet. |
4411f3b6 |
659 | |
660 | It is also planned to allow I<CHECK> to be a code reference. |
661 | |
47bfe92f |
662 | This is not yet implemented as there are design issues with what its |
663 | arguments should be and how it returns its results. |
4411f3b6 |
664 | |
665 | =over 4 |
666 | |
667 | =item Scheme 1 |
668 | |
669 | Passed remaining fragment of string being processed. |
670 | Modifies it in place to remove bytes/characters it can understand |
671 | and returns a string used to represent them. |
672 | e.g. |
673 | |
674 | sub fixup { |
675 | my $ch = substr($_[0],0,1,''); |
676 | return sprintf("\x{%02X}",ord($ch); |
677 | } |
678 | |
679 | This scheme is close to how underlying C code for Encode works, but gives |
680 | the fixup routine very little context. |
681 | |
682 | =item Scheme 2 |
683 | |
47bfe92f |
684 | Passed original string, and an index into it of the problem area, and |
685 | output string so far. Appends what it will to output string and |
686 | returns new index into original string. For example: |
4411f3b6 |
687 | |
688 | sub fixup { |
689 | # my ($s,$i,$d) = @_; |
690 | my $ch = substr($_[0],$_[1],1); |
691 | $_[2] .= sprintf("\x{%02X}",ord($ch); |
692 | return $_[1]+1; |
693 | } |
694 | |
47bfe92f |
695 | This scheme gives maximal control to the fixup routine but is more |
696 | complicated to code, and may need internals of Encode to be tweaked to |
697 | keep original string intact. |
4411f3b6 |
698 | |
699 | =item Other Schemes |
700 | |
701 | Hybrids of above. |
702 | |
703 | Multiple return values rather than in-place modifications. |
704 | |
705 | Index into the string could be pos($str) allowing s/\G...//. |
706 | |
707 | =back |
708 | |
709 | =head2 UTF-8 / utf8 |
710 | |
711 | The Unicode consortium defines the UTF-8 standard as a way of encoding |
47bfe92f |
712 | the entire Unicode repertiore as sequences of octets. This encoding is |
713 | expected to become very widespread. Perl can use this form internaly |
714 | to represent strings, so conversions to and from this form are |
715 | particularly efficient (as octets in memory do not have to change, |
716 | just the meta-data that tells Perl how to treat them). |
4411f3b6 |
717 | |
718 | =over 4 |
719 | |
720 | =item * |
721 | |
722 | $bytes = encode_utf8($string); |
723 | |
47bfe92f |
724 | The characters that comprise string are encoded in Perl's superset of UTF-8 |
4411f3b6 |
725 | and the resulting octets returned as a sequence of bytes. All possible |
726 | characters have a UTF-8 representation so this function cannot fail. |
727 | |
728 | =item * |
729 | |
730 | $string = decode_utf8($bytes [,CHECK]); |
731 | |
47bfe92f |
732 | The sequence of octets represented by $bytes is decoded from UTF-8 |
733 | into a sequence of logical characters. Not all sequences of octets |
734 | form valid UTF-8 encodings, so it is possible for this call to fail. |
735 | For CHECK see L</"Handling Malformed Data">. |
4411f3b6 |
736 | |
737 | =back |
738 | |
739 | =head2 Other Encodings of Unicode |
740 | |
47bfe92f |
741 | UTF-16 is similar to UCS-2, 16 bit or 2-byte chunks. UCS-2 can only |
7a4efbb2 |
742 | represent 0..0xFFFF, while UTF-16 has a I<surrogate pair> scheme which |
47bfe92f |
743 | allows it to cover the whole Unicode range. |
4411f3b6 |
744 | |
7a4efbb2 |
745 | Surrogates are code points set aside to encode the 0x01000..0x10FFFF |
746 | range of Unicode code points in pairs of 16-bit units. The I<high |
747 | surrogates> are the range 0xD800..0xDBFF, and the I<low surrogates> |
748 | are the range 0xDC00..0xDFFFF. The surrogate encoding is |
749 | |
750 | $hi = ($uni - 0x10000) / 0x400 + 0xD800; |
751 | $lo = ($uni - 0x10000) % 0x400 + 0xDC00; |
752 | |
753 | and the decoding is |
754 | |
755 | $uni = 0x10000 + ($hi - 0xD8000) * 0x400 + ($lo - 0xDC00); |
756 | |
8040349a |
757 | Encode implements big-endian UCS-2 aliased to "iso-10646-1" as that |
47bfe92f |
758 | happens to be the name used by that representation when used with X11 |
759 | fonts. |
4411f3b6 |
760 | |
761 | UTF-32 or UCS-4 is 32-bit or 4-byte chunks. Perl's logical characters |
762 | can be considered as being in this form without encoding. An encoding |
47bfe92f |
763 | to transfer strings in this form (e.g. to write them to a file) would |
764 | need to |
4411f3b6 |
765 | |
c079d275 |
766 | pack('L*', unpack('U*', $string)); # native |
4411f3b6 |
767 | or |
c079d275 |
768 | pack('V*', unpack('U*', $string)); # little-endian |
4411f3b6 |
769 | or |
c079d275 |
770 | pack('N*', unpack('U*', $string)); # big-endian |
4411f3b6 |
771 | |
c079d275 |
772 | depending on the endianness required. |
4411f3b6 |
773 | |
51ef4e11 |
774 | No UTF-32 encodings are implemented yet. |
4411f3b6 |
775 | |
47bfe92f |
776 | Both UCS-2 and UCS-4 style encodings can have "byte order marks" by |
777 | representing the code point 0xFFFE as the very first thing in a file. |
4411f3b6 |
778 | |
51ef4e11 |
779 | =head2 Listing available encodings |
780 | |
781 | use Encode qw(encodings); |
782 | @list = encodings(); |
783 | |
784 | Returns a list of the canonical names of the available encodings. |
785 | |
786 | =head2 Defining Aliases |
787 | |
788 | use Encode qw(define_alias); |
789 | define_alias( newName => ENCODING); |
790 | |
47bfe92f |
791 | Allows newName to be used as am alias for ENCODING. ENCODING may be |
792 | either the name of an encoding or and encoding object (as above). |
51ef4e11 |
793 | |
794 | Currently I<newName> can be specified in the following ways: |
795 | |
796 | =over 4 |
797 | |
798 | =item As a simple string. |
799 | |
800 | =item As a qr// compiled regular expression, e.g.: |
801 | |
802 | define_alias( qr/^iso8859-(\d+)$/i => '"iso-8859-$1"' ); |
803 | |
47bfe92f |
804 | In this case if I<ENCODING> is not a reference it is C<eval>-ed to |
805 | allow C<$1> etc. to be subsituted. The example is one way to names as |
806 | used in X11 font names to alias the MIME names for the iso-8859-* |
807 | family. |
51ef4e11 |
808 | |
809 | =item As a code reference, e.g.: |
810 | |
811 | define_alias( sub { return /^iso8859-(\d+)$/i ? "iso-8859-$1" : undef } , ''); |
812 | |
813 | In this case C<$_> will be set to the name that is being looked up and |
47bfe92f |
814 | I<ENCODING> is passed to the sub as its first argument. The example |
815 | is another way to names as used in X11 font names to alias the MIME |
816 | names for the iso-8859-* family. |
51ef4e11 |
817 | |
818 | =back |
819 | |
820 | =head2 Defining Encodings |
821 | |
e9692b5b |
822 | use Encode qw(define_alias); |
823 | define_encoding( $object, 'canonicalName' [,alias...]); |
51ef4e11 |
824 | |
47bfe92f |
825 | Causes I<canonicalName> to be associated with I<$object>. The object |
826 | should provide the interface described in L</"IMPLEMENTATION CLASSES"> |
827 | below. If more than two arguments are provided then additional |
828 | arguments are taken as aliases for I<$object> as for C<define_alias>. |
51ef4e11 |
829 | |
4411f3b6 |
830 | =head1 Encoding and IO |
831 | |
832 | It is very common to want to do encoding transformations when |
833 | reading or writing files, network connections, pipes etc. |
47bfe92f |
834 | If Perl is configured to use the new 'perlio' IO system then |
4411f3b6 |
835 | C<Encode> provides a "layer" (See L<perliol>) which can transform |
836 | data as it is read or written. |
837 | |
8e86646e |
838 | Here is how the blind poet would modernise the encoding: |
839 | |
42234700 |
840 | use Encode; |
8e86646e |
841 | open(my $iliad,'<:encoding(iso-8859-7)','iliad.greek'); |
842 | open(my $utf8,'>:utf8','iliad.utf8'); |
843 | my @epic = <$iliad>; |
844 | print $utf8 @epic; |
845 | close($utf8); |
846 | close($illiad); |
4411f3b6 |
847 | |
848 | In addition the new IO system can also be configured to read/write |
849 | UTF-8 encoded characters (as noted above this is efficient): |
850 | |
e9692b5b |
851 | open(my $fh,'>:utf8','anything'); |
852 | print $fh "Any \x{0021} string \N{SMILEY FACE}\n"; |
4411f3b6 |
853 | |
854 | Either of the above forms of "layer" specifications can be made the default |
855 | for a lexical scope with the C<use open ...> pragma. See L<open>. |
856 | |
857 | Once a handle is open is layers can be altered using C<binmode>. |
858 | |
47bfe92f |
859 | Without any such configuration, or if Perl itself is built using |
4411f3b6 |
860 | system's own IO, then write operations assume that file handle accepts |
861 | only I<bytes> and will C<die> if a character larger than 255 is |
862 | written to the handle. When reading, each octet from the handle |
863 | becomes a byte-in-a-character. Note that this default is the same |
47bfe92f |
864 | behaviour as bytes-only languages (including Perl before v5.6) would |
865 | have, and is sufficient to handle native 8-bit encodings |
866 | e.g. iso-8859-1, EBCDIC etc. and any legacy mechanisms for handling |
867 | other encodings and binary data. |
868 | |
869 | In other cases it is the programs responsibility to transform |
870 | characters into bytes using the API above before doing writes, and to |
871 | transform the bytes read from a handle into characters before doing |
872 | "character operations" (e.g. C<lc>, C</\W+/>, ...). |
873 | |
47bfe92f |
874 | You can also use PerlIO to convert larger amounts of data you don't |
875 | want to bring into memory. For example to convert between ISO 8859-1 |
876 | (Latin 1) and UTF-8 (or UTF-EBCDIC in EBCDIC machines): |
877 | |
e9692b5b |
878 | open(F, "<:encoding(iso-8859-1)", "data.txt") or die $!; |
879 | open(G, ">:utf8", "data.utf") or die $!; |
880 | while (<F>) { print G } |
881 | |
882 | # Could also do "print G <F>" but that would pull |
883 | # the whole file into memory just to write it out again. |
884 | |
885 | More examples: |
47bfe92f |
886 | |
e9692b5b |
887 | open(my $f, "<:encoding(cp1252)") |
888 | open(my $g, ">:encoding(iso-8859-2)") |
889 | open(my $h, ">:encoding(latin9)") # iso-8859-15 |
47bfe92f |
890 | |
891 | See L<PerlIO> for more information. |
4411f3b6 |
892 | |
1768d7eb |
893 | See also L<encoding> for how to change the default encoding of the |
d521382b |
894 | data in your script. |
1768d7eb |
895 | |
4411f3b6 |
896 | =head1 Encoding How to ... |
897 | |
898 | To do: |
899 | |
900 | =over 4 |
901 | |
902 | =item * IO with mixed content (faking iso-2020-*) |
903 | |
904 | =item * MIME's Content-Length: |
905 | |
906 | =item * UTF-8 strings in binary data. |
907 | |
47bfe92f |
908 | =item * Perl/Encode wrappers on non-Unicode XS modules. |
4411f3b6 |
909 | |
910 | =back |
911 | |
912 | =head1 Messing with Perl's Internals |
913 | |
47bfe92f |
914 | The following API uses parts of Perl's internals in the current |
915 | implementation. As such they are efficient, but may change. |
4411f3b6 |
916 | |
917 | =over 4 |
918 | |
4411f3b6 |
919 | =item * is_utf8(STRING [, CHECK]) |
920 | |
921 | [INTERNAL] Test whether the UTF-8 flag is turned on in the STRING. |
47bfe92f |
922 | If CHECK is true, also checks the data in STRING for being well-formed |
923 | UTF-8. Returns true if successful, false otherwise. |
4411f3b6 |
924 | |
925 | =item * valid_utf8(STRING) |
926 | |
47bfe92f |
927 | [INTERNAL] Test whether STRING is in a consistent state. Will return |
928 | true if string is held as bytes, or is well-formed UTF-8 and has the |
929 | UTF-8 flag on. Main reason for this routine is to allow Perl's |
930 | testsuite to check that operations have left strings in a consistent |
931 | state. |
4411f3b6 |
932 | |
933 | =item * |
934 | |
935 | _utf8_on(STRING) |
936 | |
937 | [INTERNAL] Turn on the UTF-8 flag in STRING. The data in STRING is |
938 | B<not> checked for being well-formed UTF-8. Do not use unless you |
939 | B<know> that the STRING is well-formed UTF-8. Returns the previous |
940 | state of the UTF-8 flag (so please don't test the return value as |
941 | I<not> success or failure), or C<undef> if STRING is not a string. |
942 | |
943 | =item * |
944 | |
945 | _utf8_off(STRING) |
946 | |
947 | [INTERNAL] Turn off the UTF-8 flag in STRING. Do not use frivolously. |
948 | Returns the previous state of the UTF-8 flag (so please don't test the |
949 | return value as I<not> success or failure), or C<undef> if STRING is |
950 | not a string. |
951 | |
952 | =back |
953 | |
4edaa979 |
954 | =head1 IMPLEMENTATION CLASSES |
955 | |
956 | As mentioned above encodings are (in the current implementation at least) |
957 | defined by objects. The mapping of encoding name to object is via the |
51ef4e11 |
958 | C<%encodings> hash. |
4edaa979 |
959 | |
960 | The values of the hash can currently be either strings or objects. |
961 | The string form may go away in the future. The string form occurs |
962 | when C<encodings()> has scanned C<@INC> for loadable encodings but has |
963 | not actually loaded the encoding in question. This is because the |
47bfe92f |
964 | current "loading" process is all Perl and a bit slow. |
4edaa979 |
965 | |
47bfe92f |
966 | Once an encoding is loaded then value of the hash is object which |
967 | implements the encoding. The object should provide the following |
968 | interface: |
4edaa979 |
969 | |
970 | =over 4 |
971 | |
972 | =item -E<gt>name |
973 | |
974 | Should return the string representing the canonical name of the encoding. |
975 | |
976 | =item -E<gt>new_sequence |
977 | |
47bfe92f |
978 | This is a placeholder for encodings with state. It should return an |
979 | object which implements this interface, all current implementations |
980 | return the original object. |
4edaa979 |
981 | |
982 | =item -E<gt>encode($string,$check) |
983 | |
47bfe92f |
984 | Should return the octet sequence representing I<$string>. If I<$check> |
985 | is true it should modify I<$string> in place to remove the converted |
986 | part (i.e. the whole string unless there is an error). If an error |
987 | occurs it should return the octet sequence for the fragment of string |
988 | that has been converted, and modify $string in-place to remove the |
989 | converted part leaving it starting with the problem fragment. |
4edaa979 |
990 | |
47bfe92f |
991 | If check is is false then C<encode> should make a "best effort" to |
992 | convert the string - for example by using a replacement character. |
4edaa979 |
993 | |
994 | =item -E<gt>decode($octets,$check) |
995 | |
47bfe92f |
996 | Should return the string that I<$octets> represents. If I<$check> is |
997 | true it should modify I<$octets> in place to remove the converted part |
998 | (i.e. the whole sequence unless there is an error). If an error |
999 | occurs it should return the fragment of string that has been |
1000 | converted, and modify $octets in-place to remove the converted part |
4edaa979 |
1001 | leaving it starting with the problem fragment. |
1002 | |
47bfe92f |
1003 | If check is is false then C<decode> should make a "best effort" to |
1004 | convert the string - for example by using Unicode's "\x{FFFD}" as a |
1005 | replacement character. |
4edaa979 |
1006 | |
1007 | =back |
1008 | |
47bfe92f |
1009 | It should be noted that the check behaviour is different from the |
1010 | outer public API. The logic is that the "unchecked" case is useful |
1011 | when encoding is part of a stream which may be reporting errors |
1012 | (e.g. STDERR). In such cases it is desirable to get everything |
1013 | through somehow without causing additional errors which obscure the |
1014 | original one. Also the encoding is best placed to know what the |
1015 | correct replacement character is, so if that is the desired behaviour |
1016 | then letting low level code do it is the most efficient. |
1017 | |
1018 | In contrast if check is true, the scheme above allows the encoding to |
1019 | do as much as it can and tell layer above how much that was. What is |
1020 | lacking at present is a mechanism to report what went wrong. The most |
1021 | likely interface will be an additional method call to the object, or |
1022 | perhaps (to avoid forcing per-stream objects on otherwise stateless |
1023 | encodings) and additional parameter. |
1024 | |
1025 | It is also highly desirable that encoding classes inherit from |
1026 | C<Encode::Encoding> as a base class. This allows that class to define |
1027 | additional behaviour for all encoding objects. For example built in |
1028 | Unicode, UCS-2 and UTF-8 classes use : |
51ef4e11 |
1029 | |
1030 | package Encode::MyEncoding; |
1031 | use base qw(Encode::Encoding); |
1032 | |
1033 | __PACKAGE__->Define(qw(myCanonical myAlias)); |
1034 | |
47bfe92f |
1035 | To create an object with bless {Name => ...},$class, and call |
1036 | define_encoding. They inherit their C<name> method from |
1037 | C<Encode::Encoding>. |
4edaa979 |
1038 | |
1039 | =head2 Compiled Encodings |
1040 | |
47bfe92f |
1041 | F<Encode.xs> provides a class C<Encode::XS> which provides the |
1042 | interface described above. It calls a generic octet-sequence to |
1043 | octet-sequence "engine" that is driven by tables (defined in |
1044 | F<encengine.c>). The same engine is used for both encode and |
1045 | decode. C<Encode:XS>'s C<encode> forces Perl's characters to their |
1046 | UTF-8 form and then treats them as just another multibyte |
1047 | encoding. C<Encode:XS>'s C<decode> transforms the sequence and then |
1048 | turns the UTF-8-ness flag as that is the form that the tables are |
1049 | defined to produce. For details of the engine see the comments in |
1050 | F<encengine.c>. |
1051 | |
1052 | The tables are produced by the Perl script F<compile> (the name needs |
1053 | to change so we can eventually install it somewhere). F<compile> can |
1054 | currently read two formats: |
4edaa979 |
1055 | |
1056 | =over 4 |
1057 | |
1058 | =item *.enc |
1059 | |
47bfe92f |
1060 | This is a coined format used by Tcl. It is documented in |
1061 | Encode/EncodeFormat.pod. |
4edaa979 |
1062 | |
1063 | =item *.ucm |
1064 | |
1065 | This is the semi-standard format used by IBM's ICU package. |
1066 | |
1067 | =back |
1068 | |
1069 | F<compile> can write the following forms: |
1070 | |
1071 | =over 4 |
1072 | |
1073 | =item *.ucm |
1074 | |
1075 | See above - the F<Encode/*.ucm> files provided with the distribution have |
1076 | been created from the original Tcl .enc files using this approach. |
1077 | |
1078 | =item *.c |
1079 | |
1080 | Produces tables as C data structures - this is used to build in encodings |
1081 | into F<Encode.so>/F<Encode.dll>. |
1082 | |
1083 | =item *.xs |
1084 | |
47bfe92f |
1085 | In theory this allows encodings to be stand-alone loadable Perl |
1086 | extensions. The process has not yet been tested. The plan is to use |
1087 | this approach for large East Asian encodings. |
4edaa979 |
1088 | |
1089 | =back |
1090 | |
47bfe92f |
1091 | The set of encodings built-in to F<Encode.so>/F<Encode.dll> is |
1092 | determined by F<Makefile.PL>. The current set is as follows: |
4edaa979 |
1093 | |
1094 | =over 4 |
1095 | |
1096 | =item ascii and iso-8859-* |
1097 | |
1098 | That is all the common 8-bit "western" encodings. |
1099 | |
1100 | =item IBM-1047 and two other variants of EBCDIC. |
1101 | |
47bfe92f |
1102 | These are the same variants that are supported by EBCDIC Perl as |
1103 | "native" encodings. They are included to prove "reversibility" of |
1104 | some constructs in EBCDIC Perl. |
4edaa979 |
1105 | |
1106 | =item symbol and dingbats as used by Tk on X11. |
1107 | |
47bfe92f |
1108 | (The reason Encode got started was to support Perl/Tk.) |
4edaa979 |
1109 | |
1110 | =back |
1111 | |
47bfe92f |
1112 | That set is rather ad hoc and has been driven by the needs of the |
1113 | tests rather than the needs of typical applications. It is likely |
1114 | to be rationalized. |
4edaa979 |
1115 | |
4411f3b6 |
1116 | =head1 SEE ALSO |
1117 | |
1768d7eb |
1118 | L<perlunicode>, L<perlebcdic>, L<perlfunc/open>, L<PerlIO>, L<encoding> |
4411f3b6 |
1119 | |
1120 | =cut |
1121 | |