Commit | Line | Data |
2c674647 |
1 | package Encode; |
51ef4e11 |
2 | use strict; |
2c674647 |
3 | |
b8a524e9 |
4 | our $VERSION = '0.02'; |
2c674647 |
5 | |
6 | require DynaLoader; |
7 | require Exporter; |
8 | |
51ef4e11 |
9 | our @ISA = qw(Exporter DynaLoader); |
2c674647 |
10 | |
4411f3b6 |
11 | # Public, encouraged API is exported by default |
51ef4e11 |
12 | our @EXPORT = qw ( |
4411f3b6 |
13 | encode |
14 | decode |
15 | encode_utf8 |
16 | decode_utf8 |
17 | find_encoding |
51ef4e11 |
18 | encodings |
4411f3b6 |
19 | ); |
20 | |
51ef4e11 |
21 | our @EXPORT_OK = |
2c674647 |
22 | qw( |
51ef4e11 |
23 | define_encoding |
24 | define_alias |
2c674647 |
25 | from_to |
26 | is_utf8 |
4411f3b6 |
27 | is_8bit |
28 | is_16bit |
a12c0f56 |
29 | utf8_upgrade |
30 | utf8_downgrade |
4411f3b6 |
31 | _utf8_on |
32 | _utf8_off |
2c674647 |
33 | ); |
34 | |
35 | bootstrap Encode (); |
36 | |
4411f3b6 |
37 | # Documentation moved after __END__ for speed - NI-S |
2c674647 |
38 | |
bf230f3d |
39 | use Carp; |
40 | |
51ef4e11 |
41 | # Make a %encoding package variable to allow a certain amount of cheating |
42 | our %encoding; |
43 | my @alias; # ordered matching list |
44 | my %alias; # cached known aliases |
5345d506 |
45 | |
656753f8 |
46 | sub encodings |
47 | { |
48 | my ($class) = @_; |
51ef4e11 |
49 | return keys %encoding; |
50 | } |
51 | |
52 | sub findAlias |
53 | { |
54 | my $class = shift; |
55 | local $_ = shift; |
56 | unless (exists $alias{$_}) |
656753f8 |
57 | { |
51ef4e11 |
58 | for (my $i=0; $i < @alias; $i += 2) |
656753f8 |
59 | { |
51ef4e11 |
60 | my $alias = $alias[$i]; |
61 | my $val = $alias[$i+1]; |
62 | my $new; |
5ad8ef52 |
63 | |
51ef4e11 |
64 | if (ref($alias) eq 'Regexp' && $_ =~ $alias) |
5345d506 |
65 | { |
51ef4e11 |
66 | $new = eval $val; |
67 | } |
68 | elsif (ref($alias) eq 'CODE') |
69 | { |
70 | $new = &{$alias}($val) |
71 | } |
5ad8ef52 |
72 | elsif (lc($_) eq lc($alias)) |
51ef4e11 |
73 | { |
74 | $new = $val; |
75 | } |
76 | if (defined($new)) |
77 | { |
78 | next if $new eq $_; # avoid (direct) recursion on bugs |
79 | my $enc = (ref($new)) ? $new : find_encoding($new); |
80 | if ($enc) |
5345d506 |
81 | { |
51ef4e11 |
82 | $alias{$_} = $enc; |
83 | last; |
5345d506 |
84 | } |
85 | } |
656753f8 |
86 | } |
5345d506 |
87 | } |
51ef4e11 |
88 | return $alias{$_}; |
5345d506 |
89 | } |
90 | |
51ef4e11 |
91 | sub define_alias |
5345d506 |
92 | { |
51ef4e11 |
93 | while (@_) |
5345d506 |
94 | { |
51ef4e11 |
95 | my ($alias,$name) = splice(@_,0,2); |
96 | push(@alias, $alias => $name); |
656753f8 |
97 | } |
51ef4e11 |
98 | } |
99 | |
016cb72c |
100 | # Allow variants of iso-8859-1 etc. |
d6089a2a |
101 | define_alias( qr/^iso[-_]?(\d+)[-_](\d+)$/i => '"iso-$1-$2"' ); |
016cb72c |
102 | |
103 | # Allow latin-1 style names as well |
104 | # 0 1 2 3 4 5 6 7 8 9 10 |
105 | my @latin2iso_num = ( 0, 1, 2, 3, 4, 9, 10, 13, 14, 15, 16 ); |
106 | define_alias( qr/^latin[-_]?(\d+)$/i => '"iso-8859-$latin2iso_num[$1]"' ); |
107 | |
108 | # Common names for non-latin prefered MIME names |
109 | define_alias( 'ascii' => 'US-ascii', |
110 | 'cyrillic' => 'iso-8859-5', |
111 | 'arabic' => 'iso-8859-6', |
112 | 'greek' => 'iso-8859-7', |
113 | 'hebrew' => 'iso-8859-8'); |
114 | |
51ef4e11 |
115 | define_alias( 'ibm-1047' => 'cp1047'); |
116 | |
016cb72c |
117 | # Map white space and _ to '-' |
118 | define_alias( qr/^(\S+)[\s_]+(.*)$/i => '"$1-$2"' ); |
119 | |
51ef4e11 |
120 | sub define_encoding |
121 | { |
122 | my $obj = shift; |
123 | my $name = shift; |
124 | $encoding{$name} = $obj; |
125 | my $lc = lc($name); |
126 | define_alias($lc => $obj) unless $lc eq $name; |
127 | while (@_) |
656753f8 |
128 | { |
51ef4e11 |
129 | my $alias = shift; |
130 | define_alias($alias,$obj); |
656753f8 |
131 | } |
51ef4e11 |
132 | return $obj; |
656753f8 |
133 | } |
134 | |
656753f8 |
135 | sub getEncoding |
136 | { |
137 | my ($class,$name) = @_; |
5345d506 |
138 | my $enc; |
51ef4e11 |
139 | if (exists $encoding{$name}) |
656753f8 |
140 | { |
51ef4e11 |
141 | return $encoding{$name}; |
142 | } |
143 | else |
144 | { |
145 | return $class->findAlias($name); |
656753f8 |
146 | } |
656753f8 |
147 | } |
148 | |
4411f3b6 |
149 | sub find_encoding |
150 | { |
151 | my ($name) = @_; |
152 | return __PACKAGE__->getEncoding($name); |
153 | } |
154 | |
155 | sub encode |
156 | { |
157 | my ($name,$string,$check) = @_; |
158 | my $enc = find_encoding($name); |
159 | croak("Unknown encoding '$name'") unless defined $enc; |
50d26985 |
160 | my $octets = $enc->encode($string,$check); |
4411f3b6 |
161 | return undef if ($check && length($string)); |
162 | return $octets; |
163 | } |
164 | |
165 | sub decode |
166 | { |
167 | my ($name,$octets,$check) = @_; |
168 | my $enc = find_encoding($name); |
169 | croak("Unknown encoding '$name'") unless defined $enc; |
50d26985 |
170 | my $string = $enc->decode($octets,$check); |
4411f3b6 |
171 | return undef if ($check && length($octets)); |
172 | return $string; |
173 | } |
174 | |
175 | sub from_to |
176 | { |
177 | my ($string,$from,$to,$check) = @_; |
178 | my $f = find_encoding($from); |
179 | croak("Unknown encoding '$from'") unless defined $f; |
180 | my $t = find_encoding($to); |
181 | croak("Unknown encoding '$to'") unless defined $t; |
50d26985 |
182 | my $uni = $f->decode($string,$check); |
4411f3b6 |
183 | return undef if ($check && length($string)); |
50d26985 |
184 | $string = $t->encode($uni,$check); |
4411f3b6 |
185 | return undef if ($check && length($uni)); |
186 | return length($_[0] = $string); |
187 | } |
188 | |
189 | sub encode_utf8 |
190 | { |
191 | my ($str) = @_; |
1b026014 |
192 | utf8::encode($str); |
4411f3b6 |
193 | return $str; |
194 | } |
195 | |
196 | sub decode_utf8 |
197 | { |
198 | my ($str) = @_; |
1b026014 |
199 | return undef unless utf8::decode($str); |
4411f3b6 |
200 | return $str; |
201 | } |
202 | |
50d26985 |
203 | package Encode::Encoding; |
204 | # Base class for classes which implement encodings |
4edaa979 |
205 | |
51ef4e11 |
206 | sub Define |
207 | { |
208 | my $obj = shift; |
209 | my $canonical = shift; |
210 | $obj = bless { Name => $canonical },$obj unless ref $obj; |
211 | # warn "$canonical => $obj\n"; |
212 | Encode::define_encoding($obj, $canonical, @_); |
213 | } |
214 | |
215 | sub name { shift->{'Name'} } |
216 | |
50d26985 |
217 | # Temporary legacy methods |
4edaa979 |
218 | sub toUnicode { shift->decode(@_) } |
219 | sub fromUnicode { shift->encode(@_) } |
220 | |
221 | sub new_sequence { return $_[0] } |
50d26985 |
222 | |
223 | package Encode::XS; |
224 | use base 'Encode::Encoding'; |
225 | |
5ad8ef52 |
226 | package Encode::Internal; |
50d26985 |
227 | use base 'Encode::Encoding'; |
656753f8 |
228 | |
9b37254d |
229 | # Dummy package that provides the encode interface but leaves data |
1b026014 |
230 | # as UTF-X encoded. It is here so that from_to() works. |
656753f8 |
231 | |
5ad8ef52 |
232 | __PACKAGE__->Define('Internal'); |
233 | |
234 | Encode::define_alias( 'Unicode' => 'Internal' ) if ord('A') == 65; |
656753f8 |
235 | |
50d26985 |
236 | sub decode |
a12c0f56 |
237 | { |
238 | my ($obj,$str,$chk) = @_; |
1b026014 |
239 | utf8::upgrade($str); |
a12c0f56 |
240 | $_[1] = '' if $chk; |
241 | return $str; |
242 | } |
656753f8 |
243 | |
50d26985 |
244 | *encode = \&decode; |
656753f8 |
245 | |
5ad8ef52 |
246 | package Encoding::Unicode; |
247 | use base 'Encode::Encoding'; |
248 | |
249 | __PACKAGE__->Define('Unicode') unless ord('A') == 65; |
250 | |
251 | sub decode |
252 | { |
253 | my ($obj,$str,$chk) = @_; |
254 | my $res = ''; |
255 | for (my $i = 0; $i < length($str); $i++) |
256 | { |
257 | $res .= chr(utf8::unicode_to_native(ord(substr($str,$i,1)))); |
258 | } |
259 | $_[1] = '' if $chk; |
260 | return $res; |
261 | } |
262 | |
263 | sub encode |
264 | { |
265 | my ($obj,$str,$chk) = @_; |
266 | my $res = ''; |
267 | for (my $i = 0; $i < length($str); $i++) |
268 | { |
269 | $res .= chr(utf8::native_to_unicode(ord(substr($str,$i,1)))); |
270 | } |
271 | $_[1] = '' if $chk; |
272 | return $res; |
273 | } |
274 | |
275 | |
4411f3b6 |
276 | package Encode::utf8; |
50d26985 |
277 | use base 'Encode::Encoding'; |
4411f3b6 |
278 | # package to allow long-hand |
279 | # $octets = encode( utf8 => $string ); |
280 | # |
281 | |
51ef4e11 |
282 | __PACKAGE__->Define(qw(UTF-8 utf8)); |
4411f3b6 |
283 | |
50d26985 |
284 | sub decode |
4411f3b6 |
285 | { |
286 | my ($obj,$octets,$chk) = @_; |
2a936312 |
287 | my $str = Encode::decode_utf8($octets); |
4411f3b6 |
288 | if (defined $str) |
289 | { |
290 | $_[1] = '' if $chk; |
291 | return $str; |
292 | } |
293 | return undef; |
294 | } |
295 | |
50d26985 |
296 | sub encode |
4411f3b6 |
297 | { |
298 | my ($obj,$string,$chk) = @_; |
2a936312 |
299 | my $octets = Encode::encode_utf8($string); |
4411f3b6 |
300 | $_[1] = '' if $chk; |
301 | return $octets; |
4411f3b6 |
302 | } |
303 | |
9b37254d |
304 | package Encode::iso10646_1; |
50d26985 |
305 | use base 'Encode::Encoding'; |
51ef4e11 |
306 | # Encoding is 16-bit network order Unicode (no surogates) |
9b37254d |
307 | # Used for X font encodings |
87714904 |
308 | |
8040349a |
309 | __PACKAGE__->Define(qw(UCS-2 iso-10646-1)); |
87714904 |
310 | |
50d26985 |
311 | sub decode |
87714904 |
312 | { |
313 | my ($obj,$str,$chk) = @_; |
314 | my $uni = ''; |
315 | while (length($str)) |
316 | { |
5dcbab34 |
317 | my $code = unpack('n',substr($str,0,2,'')) & 0xffff; |
87714904 |
318 | $uni .= chr($code); |
319 | } |
320 | $_[1] = $str if $chk; |
8040349a |
321 | utf8::upgrade($uni); |
87714904 |
322 | return $uni; |
323 | } |
324 | |
50d26985 |
325 | sub encode |
87714904 |
326 | { |
327 | my ($obj,$uni,$chk) = @_; |
328 | my $str = ''; |
329 | while (length($uni)) |
330 | { |
331 | my $ch = substr($uni,0,1,''); |
332 | my $x = ord($ch); |
333 | unless ($x < 32768) |
334 | { |
335 | last if ($chk); |
336 | $x = 0; |
337 | } |
5dcbab34 |
338 | $str .= pack('n',$x); |
656753f8 |
339 | } |
bf230f3d |
340 | $_[1] = $uni if $chk; |
656753f8 |
341 | return $str; |
342 | } |
343 | |
4411f3b6 |
344 | # switch back to Encode package in case we ever add AutoLoader |
345 | package Encode; |
346 | |
656753f8 |
347 | 1; |
348 | |
2a936312 |
349 | __END__ |
350 | |
4411f3b6 |
351 | =head1 NAME |
352 | |
353 | Encode - character encodings |
354 | |
355 | =head1 SYNOPSIS |
356 | |
357 | use Encode; |
358 | |
359 | =head1 DESCRIPTION |
360 | |
47bfe92f |
361 | The C<Encode> module provides the interfaces between Perl's strings |
362 | and the rest of the system. Perl strings are sequences of B<characters>. |
4411f3b6 |
363 | |
364 | The repertoire of characters that Perl can represent is at least that |
47bfe92f |
365 | defined by the Unicode Consortium. On most platforms the ordinal |
366 | values of the characters (as returned by C<ord(ch)>) is the "Unicode |
367 | codepoint" for the character (the exceptions are those platforms where |
368 | the legacy encoding is some variant of EBCDIC rather than a super-set |
369 | of ASCII - see L<perlebcdic>). |
4411f3b6 |
370 | |
371 | Traditionaly computer data has been moved around in 8-bit chunks |
372 | often called "bytes". These chunks are also known as "octets" in |
373 | networking standards. Perl is widely used to manipulate data of |
374 | many types - not only strings of characters representing human or |
375 | computer languages but also "binary" data being the machines representation |
376 | of numbers, pixels in an image - or just about anything. |
377 | |
47bfe92f |
378 | When Perl is processing "binary data" the programmer wants Perl to process |
379 | "sequences of bytes". This is not a problem for Perl - as a byte has 256 |
380 | possible values it easily fits in Perl's much larger "logical character". |
4411f3b6 |
381 | |
382 | =head2 TERMINOLOGY |
383 | |
4ac9195f |
384 | =over 4 |
4411f3b6 |
385 | |
386 | =item * |
387 | |
388 | I<character>: a character in the range 0..(2**32-1) (or more). |
47bfe92f |
389 | (What Perl's strings are made of.) |
4411f3b6 |
390 | |
391 | =item * |
392 | |
393 | I<byte>: a character in the range 0..255 |
47bfe92f |
394 | (A special case of a Perl character.) |
4411f3b6 |
395 | |
396 | =item * |
397 | |
398 | I<octet>: 8 bits of data, with ordinal values 0..255 |
47bfe92f |
399 | (Term for bytes passed to or from a non-Perl context, e.g. disk file.) |
4411f3b6 |
400 | |
401 | =back |
402 | |
403 | The marker [INTERNAL] marks Internal Implementation Details, in |
404 | general meant only for those who think they know what they are doing, |
405 | and such details may change in future releases. |
406 | |
407 | =head1 ENCODINGS |
408 | |
409 | =head2 Characteristics of an Encoding |
410 | |
411 | An encoding has a "repertoire" of characters that it can represent, |
412 | and for each representable character there is at least one sequence of |
413 | octets that represents it. |
414 | |
415 | =head2 Types of Encodings |
416 | |
417 | Encodings can be divided into the following types: |
418 | |
419 | =over 4 |
420 | |
421 | =item * Fixed length 8-bit (or less) encodings. |
422 | |
423 | Each character is a single octet so may have a repertoire of up to |
424 | 256 characters. ASCII and iso-8859-* are typical examples. |
425 | |
426 | =item * Fixed length 16-bit encodings |
427 | |
428 | Each character is two octets so may have a repertoire of up to |
47bfe92f |
429 | 65 536 characters. Unicode's UCS-2 is an example. Also used for |
4411f3b6 |
430 | encodings for East Asian languages. |
431 | |
432 | =item * Fixed length 32-bit encodings. |
433 | |
434 | Not really very "encoded" encodings. The Unicode code points |
435 | are just represented as 4-octet integers. None the less because |
436 | different architectures use different representations of integers |
437 | (so called "endian") there at least two disctinct encodings. |
438 | |
439 | =item * Multi-byte encodings |
440 | |
441 | The number of octets needed to represent a character varies. |
442 | UTF-8 is a particularly complex but regular case of a multi-byte |
443 | encoding. Several East Asian countries use a multi-byte encoding |
444 | where 1-octet is used to cover western roman characters and Asian |
445 | characters get 2-octets. |
446 | (UTF-16 is strictly a multi-byte encoding taking either 2 or 4 octets |
447 | to represent a Unicode code point.) |
448 | |
449 | =item * "Escape" encodings. |
450 | |
451 | These encodings embed "escape sequences" into the octet sequence |
452 | which describe how the following octets are to be interpreted. |
453 | The iso-2022-* family is typical. Following the escape sequence |
454 | octets are encoded by an "embedded" encoding (which will be one |
455 | of the above types) until another escape sequence switches to |
456 | a different "embedded" encoding. |
457 | |
458 | These schemes are very flexible and can handle mixed languages but are |
47bfe92f |
459 | very complex to process (and have state). No escape encodings are |
460 | implemented for Perl yet. |
4411f3b6 |
461 | |
462 | =back |
463 | |
464 | =head2 Specifying Encodings |
465 | |
466 | Encodings can be specified to the API described below in two ways: |
467 | |
468 | =over 4 |
469 | |
470 | =item 1. By name |
471 | |
47bfe92f |
472 | Encoding names are strings with characters taken from a restricted |
473 | repertoire. See L</"Encoding Names">. |
4411f3b6 |
474 | |
475 | =item 2. As an object |
476 | |
477 | Encoding objects are returned by C<find_encoding($name)>. |
478 | |
479 | =back |
480 | |
481 | =head2 Encoding Names |
482 | |
483 | Encoding names are case insensitive. White space in names is ignored. |
47bfe92f |
484 | In addition an encoding may have aliases. Each encoding has one |
485 | "canonical" name. The "canonical" name is chosen from the names of |
486 | the encoding by picking the first in the following sequence: |
4411f3b6 |
487 | |
488 | =over 4 |
489 | |
490 | =item * The MIME name as defined in IETF RFC-XXXX. |
491 | |
492 | =item * The name in the IANA registry. |
493 | |
494 | =item * The name used by the the organization that defined it. |
495 | |
496 | =back |
497 | |
498 | Because of all the alias issues, and because in the general case |
499 | encodings have state C<Encode> uses the encoding object internally |
500 | once an operation is in progress. |
501 | |
4411f3b6 |
502 | =head1 PERL ENCODING API |
503 | |
504 | =head2 Generic Encoding Interface |
505 | |
506 | =over 4 |
507 | |
508 | =item * |
509 | |
510 | $bytes = encode(ENCODING, $string[, CHECK]) |
511 | |
47bfe92f |
512 | Encodes string from Perl's internal form into I<ENCODING> and returns |
513 | a sequence of octets. For CHECK see L</"Handling Malformed Data">. |
4411f3b6 |
514 | |
515 | =item * |
516 | |
517 | $string = decode(ENCODING, $bytes[, CHECK]) |
518 | |
47bfe92f |
519 | Decode sequence of octets assumed to be in I<ENCODING> into Perl's |
520 | internal form and returns the resulting string. For CHECK see |
521 | L</"Handling Malformed Data">. |
522 | |
523 | =item * |
524 | |
525 | from_to($string, FROM_ENCODING, TO_ENCODING[, CHECK]) |
526 | |
527 | Convert the data between two encodings. How did the data in $string |
528 | originally get to be in FROM_ENCODING? Either using encode() or |
529 | through PerlIO: See L</"Encode and PerlIO">. For CHECK see |
530 | L</"Handling Malformed Data">. |
4411f3b6 |
531 | |
532 | =back |
533 | |
534 | =head2 Handling Malformed Data |
535 | |
536 | If CHECK is not set, C<undef> is returned. If the data is supposed to |
47bfe92f |
537 | be UTF-8, an optional lexical warning (category utf8) is given. If |
538 | CHECK is true but not a code reference, dies. |
4411f3b6 |
539 | |
47bfe92f |
540 | It would desirable to have a way to indicate that transform should use |
541 | the encodings "replacement character" - no such mechanism is defined yet. |
4411f3b6 |
542 | |
543 | It is also planned to allow I<CHECK> to be a code reference. |
544 | |
47bfe92f |
545 | This is not yet implemented as there are design issues with what its |
546 | arguments should be and how it returns its results. |
4411f3b6 |
547 | |
548 | =over 4 |
549 | |
550 | =item Scheme 1 |
551 | |
552 | Passed remaining fragment of string being processed. |
553 | Modifies it in place to remove bytes/characters it can understand |
554 | and returns a string used to represent them. |
555 | e.g. |
556 | |
557 | sub fixup { |
558 | my $ch = substr($_[0],0,1,''); |
559 | return sprintf("\x{%02X}",ord($ch); |
560 | } |
561 | |
562 | This scheme is close to how underlying C code for Encode works, but gives |
563 | the fixup routine very little context. |
564 | |
565 | =item Scheme 2 |
566 | |
47bfe92f |
567 | Passed original string, and an index into it of the problem area, and |
568 | output string so far. Appends what it will to output string and |
569 | returns new index into original string. For example: |
4411f3b6 |
570 | |
571 | sub fixup { |
572 | # my ($s,$i,$d) = @_; |
573 | my $ch = substr($_[0],$_[1],1); |
574 | $_[2] .= sprintf("\x{%02X}",ord($ch); |
575 | return $_[1]+1; |
576 | } |
577 | |
47bfe92f |
578 | This scheme gives maximal control to the fixup routine but is more |
579 | complicated to code, and may need internals of Encode to be tweaked to |
580 | keep original string intact. |
4411f3b6 |
581 | |
582 | =item Other Schemes |
583 | |
584 | Hybrids of above. |
585 | |
586 | Multiple return values rather than in-place modifications. |
587 | |
588 | Index into the string could be pos($str) allowing s/\G...//. |
589 | |
590 | =back |
591 | |
592 | =head2 UTF-8 / utf8 |
593 | |
594 | The Unicode consortium defines the UTF-8 standard as a way of encoding |
47bfe92f |
595 | the entire Unicode repertiore as sequences of octets. This encoding is |
596 | expected to become very widespread. Perl can use this form internaly |
597 | to represent strings, so conversions to and from this form are |
598 | particularly efficient (as octets in memory do not have to change, |
599 | just the meta-data that tells Perl how to treat them). |
4411f3b6 |
600 | |
601 | =over 4 |
602 | |
603 | =item * |
604 | |
605 | $bytes = encode_utf8($string); |
606 | |
47bfe92f |
607 | The characters that comprise string are encoded in Perl's superset of UTF-8 |
4411f3b6 |
608 | and the resulting octets returned as a sequence of bytes. All possible |
609 | characters have a UTF-8 representation so this function cannot fail. |
610 | |
611 | =item * |
612 | |
613 | $string = decode_utf8($bytes [,CHECK]); |
614 | |
47bfe92f |
615 | The sequence of octets represented by $bytes is decoded from UTF-8 |
616 | into a sequence of logical characters. Not all sequences of octets |
617 | form valid UTF-8 encodings, so it is possible for this call to fail. |
618 | For CHECK see L</"Handling Malformed Data">. |
4411f3b6 |
619 | |
620 | =back |
621 | |
622 | =head2 Other Encodings of Unicode |
623 | |
47bfe92f |
624 | UTF-16 is similar to UCS-2, 16 bit or 2-byte chunks. UCS-2 can only |
625 | represent 0..0xFFFF, while UTF-16 has a "surrogate pair" scheme which |
626 | allows it to cover the whole Unicode range. |
4411f3b6 |
627 | |
8040349a |
628 | Encode implements big-endian UCS-2 aliased to "iso-10646-1" as that |
47bfe92f |
629 | happens to be the name used by that representation when used with X11 |
630 | fonts. |
4411f3b6 |
631 | |
632 | UTF-32 or UCS-4 is 32-bit or 4-byte chunks. Perl's logical characters |
633 | can be considered as being in this form without encoding. An encoding |
47bfe92f |
634 | to transfer strings in this form (e.g. to write them to a file) would |
635 | need to |
4411f3b6 |
636 | |
637 | pack('L',map(chr($_),split(//,$string))); # native |
638 | or |
639 | pack('V',map(chr($_),split(//,$string))); # little-endian |
640 | or |
641 | pack('N',map(chr($_),split(//,$string))); # big-endian |
642 | |
643 | depending on the endian required. |
644 | |
51ef4e11 |
645 | No UTF-32 encodings are implemented yet. |
4411f3b6 |
646 | |
47bfe92f |
647 | Both UCS-2 and UCS-4 style encodings can have "byte order marks" by |
648 | representing the code point 0xFFFE as the very first thing in a file. |
4411f3b6 |
649 | |
51ef4e11 |
650 | =head2 Listing available encodings |
651 | |
652 | use Encode qw(encodings); |
653 | @list = encodings(); |
654 | |
655 | Returns a list of the canonical names of the available encodings. |
656 | |
657 | =head2 Defining Aliases |
658 | |
659 | use Encode qw(define_alias); |
660 | define_alias( newName => ENCODING); |
661 | |
47bfe92f |
662 | Allows newName to be used as am alias for ENCODING. ENCODING may be |
663 | either the name of an encoding or and encoding object (as above). |
51ef4e11 |
664 | |
665 | Currently I<newName> can be specified in the following ways: |
666 | |
667 | =over 4 |
668 | |
669 | =item As a simple string. |
670 | |
671 | =item As a qr// compiled regular expression, e.g.: |
672 | |
673 | define_alias( qr/^iso8859-(\d+)$/i => '"iso-8859-$1"' ); |
674 | |
47bfe92f |
675 | In this case if I<ENCODING> is not a reference it is C<eval>-ed to |
676 | allow C<$1> etc. to be subsituted. The example is one way to names as |
677 | used in X11 font names to alias the MIME names for the iso-8859-* |
678 | family. |
51ef4e11 |
679 | |
680 | =item As a code reference, e.g.: |
681 | |
682 | define_alias( sub { return /^iso8859-(\d+)$/i ? "iso-8859-$1" : undef } , ''); |
683 | |
684 | In this case C<$_> will be set to the name that is being looked up and |
47bfe92f |
685 | I<ENCODING> is passed to the sub as its first argument. The example |
686 | is another way to names as used in X11 font names to alias the MIME |
687 | names for the iso-8859-* family. |
51ef4e11 |
688 | |
689 | =back |
690 | |
691 | =head2 Defining Encodings |
692 | |
693 | use Encode qw(define_alias); |
694 | define_encoding( $object, 'canonicalName' [,alias...]); |
695 | |
47bfe92f |
696 | Causes I<canonicalName> to be associated with I<$object>. The object |
697 | should provide the interface described in L</"IMPLEMENTATION CLASSES"> |
698 | below. If more than two arguments are provided then additional |
699 | arguments are taken as aliases for I<$object> as for C<define_alias>. |
51ef4e11 |
700 | |
4411f3b6 |
701 | =head1 Encoding and IO |
702 | |
703 | It is very common to want to do encoding transformations when |
704 | reading or writing files, network connections, pipes etc. |
47bfe92f |
705 | If Perl is configured to use the new 'perlio' IO system then |
4411f3b6 |
706 | C<Encode> provides a "layer" (See L<perliol>) which can transform |
707 | data as it is read or written. |
708 | |
51ef4e11 |
709 | open(my $ilyad,'>:encoding(iso-8859-7)','ilyad.greek'); |
4411f3b6 |
710 | print $ilyad @epic; |
711 | |
712 | In addition the new IO system can also be configured to read/write |
713 | UTF-8 encoded characters (as noted above this is efficient): |
714 | |
715 | open(my $fh,'>:utf8','anything'); |
716 | print $fh "Any \x{0021} string \N{SMILEY FACE}\n"; |
717 | |
718 | Either of the above forms of "layer" specifications can be made the default |
719 | for a lexical scope with the C<use open ...> pragma. See L<open>. |
720 | |
721 | Once a handle is open is layers can be altered using C<binmode>. |
722 | |
47bfe92f |
723 | Without any such configuration, or if Perl itself is built using |
4411f3b6 |
724 | system's own IO, then write operations assume that file handle accepts |
725 | only I<bytes> and will C<die> if a character larger than 255 is |
726 | written to the handle. When reading, each octet from the handle |
727 | becomes a byte-in-a-character. Note that this default is the same |
47bfe92f |
728 | behaviour as bytes-only languages (including Perl before v5.6) would |
729 | have, and is sufficient to handle native 8-bit encodings |
730 | e.g. iso-8859-1, EBCDIC etc. and any legacy mechanisms for handling |
731 | other encodings and binary data. |
732 | |
733 | In other cases it is the programs responsibility to transform |
734 | characters into bytes using the API above before doing writes, and to |
735 | transform the bytes read from a handle into characters before doing |
736 | "character operations" (e.g. C<lc>, C</\W+/>, ...). |
737 | |
738 | =head1 Encode and PerlIO |
739 | |
740 | The PerlIO layer (new since Perl 5.7) can be used to automatically |
741 | convert the data being read in or written out to be converted from |
742 | some encoding into Perl's internal encoding or from Perl's internal |
743 | encoding into some other encoding. |
744 | |
745 | Examples: |
4411f3b6 |
746 | |
47bfe92f |
747 | open(my $f, "<:encoding(cp1252)") |
748 | |
749 | open(my $g, ">:encoding(iso-8859-1)") |
750 | |
751 | You can also use PerlIO to convert larger amounts of data you don't |
752 | want to bring into memory. For example to convert between ISO 8859-1 |
753 | (Latin 1) and UTF-8 (or UTF-EBCDIC in EBCDIC machines): |
754 | |
755 | open(F, "<:encoding(iso-8859-1)", "data.txt") or die $!; |
756 | open(G, ">:utf8", "data.utf") or die $!; |
757 | while (<F>) { print G } |
758 | |
759 | # Could do "print G <F>" but that would pull |
760 | # the whole file into memory just to write it out again. |
761 | |
762 | See L<PerlIO> for more information. |
4411f3b6 |
763 | |
764 | =head1 Encoding How to ... |
765 | |
766 | To do: |
767 | |
768 | =over 4 |
769 | |
770 | =item * IO with mixed content (faking iso-2020-*) |
771 | |
772 | =item * MIME's Content-Length: |
773 | |
774 | =item * UTF-8 strings in binary data. |
775 | |
47bfe92f |
776 | =item * Perl/Encode wrappers on non-Unicode XS modules. |
4411f3b6 |
777 | |
778 | =back |
779 | |
780 | =head1 Messing with Perl's Internals |
781 | |
47bfe92f |
782 | The following API uses parts of Perl's internals in the current |
783 | implementation. As such they are efficient, but may change. |
4411f3b6 |
784 | |
785 | =over 4 |
786 | |
4411f3b6 |
787 | =item * is_utf8(STRING [, CHECK]) |
788 | |
789 | [INTERNAL] Test whether the UTF-8 flag is turned on in the STRING. |
47bfe92f |
790 | If CHECK is true, also checks the data in STRING for being well-formed |
791 | UTF-8. Returns true if successful, false otherwise. |
4411f3b6 |
792 | |
793 | =item * valid_utf8(STRING) |
794 | |
47bfe92f |
795 | [INTERNAL] Test whether STRING is in a consistent state. Will return |
796 | true if string is held as bytes, or is well-formed UTF-8 and has the |
797 | UTF-8 flag on. Main reason for this routine is to allow Perl's |
798 | testsuite to check that operations have left strings in a consistent |
799 | state. |
4411f3b6 |
800 | |
801 | =item * |
802 | |
803 | _utf8_on(STRING) |
804 | |
805 | [INTERNAL] Turn on the UTF-8 flag in STRING. The data in STRING is |
806 | B<not> checked for being well-formed UTF-8. Do not use unless you |
807 | B<know> that the STRING is well-formed UTF-8. Returns the previous |
808 | state of the UTF-8 flag (so please don't test the return value as |
809 | I<not> success or failure), or C<undef> if STRING is not a string. |
810 | |
811 | =item * |
812 | |
813 | _utf8_off(STRING) |
814 | |
815 | [INTERNAL] Turn off the UTF-8 flag in STRING. Do not use frivolously. |
816 | Returns the previous state of the UTF-8 flag (so please don't test the |
817 | return value as I<not> success or failure), or C<undef> if STRING is |
818 | not a string. |
819 | |
820 | =back |
821 | |
4edaa979 |
822 | =head1 IMPLEMENTATION CLASSES |
823 | |
824 | As mentioned above encodings are (in the current implementation at least) |
825 | defined by objects. The mapping of encoding name to object is via the |
51ef4e11 |
826 | C<%encodings> hash. |
4edaa979 |
827 | |
828 | The values of the hash can currently be either strings or objects. |
829 | The string form may go away in the future. The string form occurs |
830 | when C<encodings()> has scanned C<@INC> for loadable encodings but has |
831 | not actually loaded the encoding in question. This is because the |
47bfe92f |
832 | current "loading" process is all Perl and a bit slow. |
4edaa979 |
833 | |
47bfe92f |
834 | Once an encoding is loaded then value of the hash is object which |
835 | implements the encoding. The object should provide the following |
836 | interface: |
4edaa979 |
837 | |
838 | =over 4 |
839 | |
840 | =item -E<gt>name |
841 | |
842 | Should return the string representing the canonical name of the encoding. |
843 | |
844 | =item -E<gt>new_sequence |
845 | |
47bfe92f |
846 | This is a placeholder for encodings with state. It should return an |
847 | object which implements this interface, all current implementations |
848 | return the original object. |
4edaa979 |
849 | |
850 | =item -E<gt>encode($string,$check) |
851 | |
47bfe92f |
852 | Should return the octet sequence representing I<$string>. If I<$check> |
853 | is true it should modify I<$string> in place to remove the converted |
854 | part (i.e. the whole string unless there is an error). If an error |
855 | occurs it should return the octet sequence for the fragment of string |
856 | that has been converted, and modify $string in-place to remove the |
857 | converted part leaving it starting with the problem fragment. |
4edaa979 |
858 | |
47bfe92f |
859 | If check is is false then C<encode> should make a "best effort" to |
860 | convert the string - for example by using a replacement character. |
4edaa979 |
861 | |
862 | =item -E<gt>decode($octets,$check) |
863 | |
47bfe92f |
864 | Should return the string that I<$octets> represents. If I<$check> is |
865 | true it should modify I<$octets> in place to remove the converted part |
866 | (i.e. the whole sequence unless there is an error). If an error |
867 | occurs it should return the fragment of string that has been |
868 | converted, and modify $octets in-place to remove the converted part |
4edaa979 |
869 | leaving it starting with the problem fragment. |
870 | |
47bfe92f |
871 | If check is is false then C<decode> should make a "best effort" to |
872 | convert the string - for example by using Unicode's "\x{FFFD}" as a |
873 | replacement character. |
4edaa979 |
874 | |
875 | =back |
876 | |
47bfe92f |
877 | It should be noted that the check behaviour is different from the |
878 | outer public API. The logic is that the "unchecked" case is useful |
879 | when encoding is part of a stream which may be reporting errors |
880 | (e.g. STDERR). In such cases it is desirable to get everything |
881 | through somehow without causing additional errors which obscure the |
882 | original one. Also the encoding is best placed to know what the |
883 | correct replacement character is, so if that is the desired behaviour |
884 | then letting low level code do it is the most efficient. |
885 | |
886 | In contrast if check is true, the scheme above allows the encoding to |
887 | do as much as it can and tell layer above how much that was. What is |
888 | lacking at present is a mechanism to report what went wrong. The most |
889 | likely interface will be an additional method call to the object, or |
890 | perhaps (to avoid forcing per-stream objects on otherwise stateless |
891 | encodings) and additional parameter. |
892 | |
893 | It is also highly desirable that encoding classes inherit from |
894 | C<Encode::Encoding> as a base class. This allows that class to define |
895 | additional behaviour for all encoding objects. For example built in |
896 | Unicode, UCS-2 and UTF-8 classes use : |
51ef4e11 |
897 | |
898 | package Encode::MyEncoding; |
899 | use base qw(Encode::Encoding); |
900 | |
901 | __PACKAGE__->Define(qw(myCanonical myAlias)); |
902 | |
47bfe92f |
903 | To create an object with bless {Name => ...},$class, and call |
904 | define_encoding. They inherit their C<name> method from |
905 | C<Encode::Encoding>. |
4edaa979 |
906 | |
907 | =head2 Compiled Encodings |
908 | |
47bfe92f |
909 | F<Encode.xs> provides a class C<Encode::XS> which provides the |
910 | interface described above. It calls a generic octet-sequence to |
911 | octet-sequence "engine" that is driven by tables (defined in |
912 | F<encengine.c>). The same engine is used for both encode and |
913 | decode. C<Encode:XS>'s C<encode> forces Perl's characters to their |
914 | UTF-8 form and then treats them as just another multibyte |
915 | encoding. C<Encode:XS>'s C<decode> transforms the sequence and then |
916 | turns the UTF-8-ness flag as that is the form that the tables are |
917 | defined to produce. For details of the engine see the comments in |
918 | F<encengine.c>. |
919 | |
920 | The tables are produced by the Perl script F<compile> (the name needs |
921 | to change so we can eventually install it somewhere). F<compile> can |
922 | currently read two formats: |
4edaa979 |
923 | |
924 | =over 4 |
925 | |
926 | =item *.enc |
927 | |
47bfe92f |
928 | This is a coined format used by Tcl. It is documented in |
929 | Encode/EncodeFormat.pod. |
4edaa979 |
930 | |
931 | =item *.ucm |
932 | |
933 | This is the semi-standard format used by IBM's ICU package. |
934 | |
935 | =back |
936 | |
937 | F<compile> can write the following forms: |
938 | |
939 | =over 4 |
940 | |
941 | =item *.ucm |
942 | |
943 | See above - the F<Encode/*.ucm> files provided with the distribution have |
944 | been created from the original Tcl .enc files using this approach. |
945 | |
946 | =item *.c |
947 | |
948 | Produces tables as C data structures - this is used to build in encodings |
949 | into F<Encode.so>/F<Encode.dll>. |
950 | |
951 | =item *.xs |
952 | |
47bfe92f |
953 | In theory this allows encodings to be stand-alone loadable Perl |
954 | extensions. The process has not yet been tested. The plan is to use |
955 | this approach for large East Asian encodings. |
4edaa979 |
956 | |
957 | =back |
958 | |
47bfe92f |
959 | The set of encodings built-in to F<Encode.so>/F<Encode.dll> is |
960 | determined by F<Makefile.PL>. The current set is as follows: |
4edaa979 |
961 | |
962 | =over 4 |
963 | |
964 | =item ascii and iso-8859-* |
965 | |
966 | That is all the common 8-bit "western" encodings. |
967 | |
968 | =item IBM-1047 and two other variants of EBCDIC. |
969 | |
47bfe92f |
970 | These are the same variants that are supported by EBCDIC Perl as |
971 | "native" encodings. They are included to prove "reversibility" of |
972 | some constructs in EBCDIC Perl. |
4edaa979 |
973 | |
974 | =item symbol and dingbats as used by Tk on X11. |
975 | |
47bfe92f |
976 | (The reason Encode got started was to support Perl/Tk.) |
4edaa979 |
977 | |
978 | =back |
979 | |
47bfe92f |
980 | That set is rather ad hoc and has been driven by the needs of the |
981 | tests rather than the needs of typical applications. It is likely |
982 | to be rationalized. |
4edaa979 |
983 | |
4411f3b6 |
984 | =head1 SEE ALSO |
985 | |
47bfe92f |
986 | L<perlunicode>, L<perlebcdic>, L<perlfunc/open>, L<PerlIO> |
4411f3b6 |
987 | |
988 | =cut |
989 | |