Commit | Line | Data |
2c674647 |
1 | package Encode; |
51ef4e11 |
2 | use strict; |
2c674647 |
3 | |
b8a524e9 |
4 | our $VERSION = '0.02'; |
2c674647 |
5 | |
6 | require DynaLoader; |
7 | require Exporter; |
8 | |
51ef4e11 |
9 | our @ISA = qw(Exporter DynaLoader); |
2c674647 |
10 | |
4411f3b6 |
11 | # Public, encouraged API is exported by default |
51ef4e11 |
12 | our @EXPORT = qw ( |
4411f3b6 |
13 | encode |
14 | decode |
15 | encode_utf8 |
16 | decode_utf8 |
17 | find_encoding |
51ef4e11 |
18 | encodings |
4411f3b6 |
19 | ); |
20 | |
51ef4e11 |
21 | our @EXPORT_OK = |
2c674647 |
22 | qw( |
51ef4e11 |
23 | define_encoding |
24 | define_alias |
2c674647 |
25 | from_to |
26 | is_utf8 |
4411f3b6 |
27 | is_8bit |
28 | is_16bit |
a12c0f56 |
29 | utf8_upgrade |
30 | utf8_downgrade |
4411f3b6 |
31 | _utf8_on |
32 | _utf8_off |
2c674647 |
33 | ); |
34 | |
35 | bootstrap Encode (); |
36 | |
4411f3b6 |
37 | # Documentation moved after __END__ for speed - NI-S |
2c674647 |
38 | |
bf230f3d |
39 | use Carp; |
40 | |
51ef4e11 |
41 | # Make a %encoding package variable to allow a certain amount of cheating |
42 | our %encoding; |
43 | my @alias; # ordered matching list |
44 | my %alias; # cached known aliases |
6d6a7c8d |
45 | # 0 1 2 3 4 5 6 7 8 9 10 |
46 | our @latin2iso_num = ( 0, 1, 2, 3, 4, 9, 10, 13, 14, 15, 16 ); |
47 | |
5345d506 |
48 | |
656753f8 |
49 | sub encodings |
50 | { |
51 | my ($class) = @_; |
51ef4e11 |
52 | return keys %encoding; |
53 | } |
54 | |
55 | sub findAlias |
56 | { |
57 | my $class = shift; |
58 | local $_ = shift; |
59 | unless (exists $alias{$_}) |
656753f8 |
60 | { |
51ef4e11 |
61 | for (my $i=0; $i < @alias; $i += 2) |
656753f8 |
62 | { |
51ef4e11 |
63 | my $alias = $alias[$i]; |
64 | my $val = $alias[$i+1]; |
65 | my $new; |
66 | if (ref($alias) eq 'Regexp' && $_ =~ $alias) |
5345d506 |
67 | { |
51ef4e11 |
68 | $new = eval $val; |
69 | } |
70 | elsif (ref($alias) eq 'CODE') |
71 | { |
72 | $new = &{$alias}($val) |
73 | } |
5ad8ef52 |
74 | elsif (lc($_) eq lc($alias)) |
51ef4e11 |
75 | { |
76 | $new = $val; |
77 | } |
78 | if (defined($new)) |
79 | { |
80 | next if $new eq $_; # avoid (direct) recursion on bugs |
81 | my $enc = (ref($new)) ? $new : find_encoding($new); |
82 | if ($enc) |
5345d506 |
83 | { |
51ef4e11 |
84 | $alias{$_} = $enc; |
85 | last; |
5345d506 |
86 | } |
87 | } |
656753f8 |
88 | } |
5345d506 |
89 | } |
51ef4e11 |
90 | return $alias{$_}; |
5345d506 |
91 | } |
92 | |
51ef4e11 |
93 | sub define_alias |
5345d506 |
94 | { |
51ef4e11 |
95 | while (@_) |
5345d506 |
96 | { |
51ef4e11 |
97 | my ($alias,$name) = splice(@_,0,2); |
98 | push(@alias, $alias => $name); |
656753f8 |
99 | } |
51ef4e11 |
100 | } |
101 | |
016cb72c |
102 | # Allow variants of iso-8859-1 etc. |
d6089a2a |
103 | define_alias( qr/^iso[-_]?(\d+)[-_](\d+)$/i => '"iso-$1-$2"' ); |
016cb72c |
104 | |
7faf300d |
105 | # At least HP-UX has these. |
106 | define_alias( qr/^iso8859(\d+)$/i => '"iso-8859-$1"' ); |
107 | |
58d53262 |
108 | # This is a font issue, not an encoding issue. |
109 | # (The currency symbol of the Latin 1 upper half |
110 | # has been redefined as the euro symbol.) |
111 | define_alias( qr/^(.+)\@euro$/i => '"$1"' ); |
112 | |
016cb72c |
113 | # Allow latin-1 style names as well |
7faf300d |
114 | define_alias( qr/^(?:iso[-_]?)?latin[-_]?(\d+)$/i => '"iso-8859-$latin2iso_num[$1]"' ); |
016cb72c |
115 | |
116 | # Common names for non-latin prefered MIME names |
117 | define_alias( 'ascii' => 'US-ascii', |
118 | 'cyrillic' => 'iso-8859-5', |
119 | 'arabic' => 'iso-8859-6', |
120 | 'greek' => 'iso-8859-7', |
121 | 'hebrew' => 'iso-8859-8'); |
122 | |
7faf300d |
123 | # At least AIX has IBM-NNN (surprisingly...) instead of cpNNN. |
124 | define_alias( qr/^ibm[-_]?(\d\d\d\d?)$/i => '"cp$1"'); |
125 | |
58d53262 |
126 | # Standardize on the dashed versions. |
127 | define_alias( qr/^utf8$/i => 'utf-8' ); |
7faf300d |
128 | define_alias( qr/^koi8r$/i => 'koi8-r' ); |
51ef4e11 |
129 | |
58d53262 |
130 | # TODO: the HP-UX '8' encodings: arabic8 greek8 hebrew8 roman8 turkish8 |
131 | # TODO: the Thai Encoding tis620 |
132 | # TODO: the Chinese Encoding gb18030 |
133 | # TODO: what is the Japanese 'ujis' encoding seen in some Linuxes? |
134 | |
016cb72c |
135 | # Map white space and _ to '-' |
136 | define_alias( qr/^(\S+)[\s_]+(.*)$/i => '"$1-$2"' ); |
137 | |
51ef4e11 |
138 | sub define_encoding |
139 | { |
140 | my $obj = shift; |
141 | my $name = shift; |
142 | $encoding{$name} = $obj; |
143 | my $lc = lc($name); |
144 | define_alias($lc => $obj) unless $lc eq $name; |
145 | while (@_) |
656753f8 |
146 | { |
51ef4e11 |
147 | my $alias = shift; |
148 | define_alias($alias,$obj); |
656753f8 |
149 | } |
51ef4e11 |
150 | return $obj; |
656753f8 |
151 | } |
152 | |
656753f8 |
153 | sub getEncoding |
154 | { |
155 | my ($class,$name) = @_; |
5345d506 |
156 | my $enc; |
0f43fc90 |
157 | if (ref($name) && $name->can('new_sequence')) |
158 | { |
159 | return $name; |
160 | } |
51ef4e11 |
161 | if (exists $encoding{$name}) |
656753f8 |
162 | { |
51ef4e11 |
163 | return $encoding{$name}; |
164 | } |
165 | else |
166 | { |
167 | return $class->findAlias($name); |
656753f8 |
168 | } |
656753f8 |
169 | } |
170 | |
4411f3b6 |
171 | sub find_encoding |
172 | { |
173 | my ($name) = @_; |
174 | return __PACKAGE__->getEncoding($name); |
175 | } |
176 | |
177 | sub encode |
178 | { |
179 | my ($name,$string,$check) = @_; |
180 | my $enc = find_encoding($name); |
181 | croak("Unknown encoding '$name'") unless defined $enc; |
50d26985 |
182 | my $octets = $enc->encode($string,$check); |
4411f3b6 |
183 | return undef if ($check && length($string)); |
184 | return $octets; |
185 | } |
186 | |
187 | sub decode |
188 | { |
189 | my ($name,$octets,$check) = @_; |
190 | my $enc = find_encoding($name); |
191 | croak("Unknown encoding '$name'") unless defined $enc; |
50d26985 |
192 | my $string = $enc->decode($octets,$check); |
4411f3b6 |
193 | return undef if ($check && length($octets)); |
194 | return $string; |
195 | } |
196 | |
197 | sub from_to |
198 | { |
199 | my ($string,$from,$to,$check) = @_; |
200 | my $f = find_encoding($from); |
201 | croak("Unknown encoding '$from'") unless defined $f; |
202 | my $t = find_encoding($to); |
203 | croak("Unknown encoding '$to'") unless defined $t; |
50d26985 |
204 | my $uni = $f->decode($string,$check); |
4411f3b6 |
205 | return undef if ($check && length($string)); |
50d26985 |
206 | $string = $t->encode($uni,$check); |
4411f3b6 |
207 | return undef if ($check && length($uni)); |
208 | return length($_[0] = $string); |
209 | } |
210 | |
211 | sub encode_utf8 |
212 | { |
213 | my ($str) = @_; |
1b026014 |
214 | utf8::encode($str); |
4411f3b6 |
215 | return $str; |
216 | } |
217 | |
218 | sub decode_utf8 |
219 | { |
220 | my ($str) = @_; |
1b026014 |
221 | return undef unless utf8::decode($str); |
4411f3b6 |
222 | return $str; |
223 | } |
224 | |
50d26985 |
225 | package Encode::Encoding; |
226 | # Base class for classes which implement encodings |
4edaa979 |
227 | |
51ef4e11 |
228 | sub Define |
229 | { |
230 | my $obj = shift; |
231 | my $canonical = shift; |
232 | $obj = bless { Name => $canonical },$obj unless ref $obj; |
233 | # warn "$canonical => $obj\n"; |
234 | Encode::define_encoding($obj, $canonical, @_); |
235 | } |
236 | |
237 | sub name { shift->{'Name'} } |
238 | |
50d26985 |
239 | # Temporary legacy methods |
4edaa979 |
240 | sub toUnicode { shift->decode(@_) } |
241 | sub fromUnicode { shift->encode(@_) } |
242 | |
243 | sub new_sequence { return $_[0] } |
50d26985 |
244 | |
245 | package Encode::XS; |
246 | use base 'Encode::Encoding'; |
247 | |
5ad8ef52 |
248 | package Encode::Internal; |
50d26985 |
249 | use base 'Encode::Encoding'; |
656753f8 |
250 | |
9b37254d |
251 | # Dummy package that provides the encode interface but leaves data |
1b026014 |
252 | # as UTF-X encoded. It is here so that from_to() works. |
656753f8 |
253 | |
5ad8ef52 |
254 | __PACKAGE__->Define('Internal'); |
255 | |
256 | Encode::define_alias( 'Unicode' => 'Internal' ) if ord('A') == 65; |
656753f8 |
257 | |
50d26985 |
258 | sub decode |
a12c0f56 |
259 | { |
260 | my ($obj,$str,$chk) = @_; |
1b026014 |
261 | utf8::upgrade($str); |
a12c0f56 |
262 | $_[1] = '' if $chk; |
263 | return $str; |
264 | } |
656753f8 |
265 | |
50d26985 |
266 | *encode = \&decode; |
656753f8 |
267 | |
5ad8ef52 |
268 | package Encoding::Unicode; |
269 | use base 'Encode::Encoding'; |
270 | |
271 | __PACKAGE__->Define('Unicode') unless ord('A') == 65; |
272 | |
273 | sub decode |
274 | { |
275 | my ($obj,$str,$chk) = @_; |
276 | my $res = ''; |
277 | for (my $i = 0; $i < length($str); $i++) |
278 | { |
279 | $res .= chr(utf8::unicode_to_native(ord(substr($str,$i,1)))); |
280 | } |
281 | $_[1] = '' if $chk; |
282 | return $res; |
283 | } |
284 | |
285 | sub encode |
286 | { |
287 | my ($obj,$str,$chk) = @_; |
288 | my $res = ''; |
289 | for (my $i = 0; $i < length($str); $i++) |
290 | { |
291 | $res .= chr(utf8::native_to_unicode(ord(substr($str,$i,1)))); |
292 | } |
293 | $_[1] = '' if $chk; |
294 | return $res; |
295 | } |
296 | |
297 | |
4411f3b6 |
298 | package Encode::utf8; |
50d26985 |
299 | use base 'Encode::Encoding'; |
4411f3b6 |
300 | # package to allow long-hand |
301 | # $octets = encode( utf8 => $string ); |
302 | # |
303 | |
51ef4e11 |
304 | __PACKAGE__->Define(qw(UTF-8 utf8)); |
4411f3b6 |
305 | |
50d26985 |
306 | sub decode |
4411f3b6 |
307 | { |
308 | my ($obj,$octets,$chk) = @_; |
2a936312 |
309 | my $str = Encode::decode_utf8($octets); |
4411f3b6 |
310 | if (defined $str) |
311 | { |
312 | $_[1] = '' if $chk; |
313 | return $str; |
314 | } |
315 | return undef; |
316 | } |
317 | |
50d26985 |
318 | sub encode |
4411f3b6 |
319 | { |
320 | my ($obj,$string,$chk) = @_; |
2a936312 |
321 | my $octets = Encode::encode_utf8($string); |
4411f3b6 |
322 | $_[1] = '' if $chk; |
323 | return $octets; |
4411f3b6 |
324 | } |
325 | |
9b37254d |
326 | package Encode::iso10646_1; |
50d26985 |
327 | use base 'Encode::Encoding'; |
51ef4e11 |
328 | # Encoding is 16-bit network order Unicode (no surogates) |
9b37254d |
329 | # Used for X font encodings |
87714904 |
330 | |
8040349a |
331 | __PACKAGE__->Define(qw(UCS-2 iso-10646-1)); |
87714904 |
332 | |
50d26985 |
333 | sub decode |
87714904 |
334 | { |
335 | my ($obj,$str,$chk) = @_; |
336 | my $uni = ''; |
337 | while (length($str)) |
338 | { |
5dcbab34 |
339 | my $code = unpack('n',substr($str,0,2,'')) & 0xffff; |
87714904 |
340 | $uni .= chr($code); |
341 | } |
342 | $_[1] = $str if $chk; |
8040349a |
343 | utf8::upgrade($uni); |
87714904 |
344 | return $uni; |
345 | } |
346 | |
50d26985 |
347 | sub encode |
87714904 |
348 | { |
349 | my ($obj,$uni,$chk) = @_; |
350 | my $str = ''; |
351 | while (length($uni)) |
352 | { |
353 | my $ch = substr($uni,0,1,''); |
354 | my $x = ord($ch); |
355 | unless ($x < 32768) |
356 | { |
357 | last if ($chk); |
358 | $x = 0; |
359 | } |
5dcbab34 |
360 | $str .= pack('n',$x); |
656753f8 |
361 | } |
bf230f3d |
362 | $_[1] = $uni if $chk; |
656753f8 |
363 | return $str; |
364 | } |
365 | |
4411f3b6 |
366 | # switch back to Encode package in case we ever add AutoLoader |
367 | package Encode; |
368 | |
656753f8 |
369 | 1; |
370 | |
2a936312 |
371 | __END__ |
372 | |
4411f3b6 |
373 | =head1 NAME |
374 | |
375 | Encode - character encodings |
376 | |
377 | =head1 SYNOPSIS |
378 | |
379 | use Encode; |
380 | |
381 | =head1 DESCRIPTION |
382 | |
47bfe92f |
383 | The C<Encode> module provides the interfaces between Perl's strings |
384 | and the rest of the system. Perl strings are sequences of B<characters>. |
4411f3b6 |
385 | |
386 | The repertoire of characters that Perl can represent is at least that |
47bfe92f |
387 | defined by the Unicode Consortium. On most platforms the ordinal |
388 | values of the characters (as returned by C<ord(ch)>) is the "Unicode |
389 | codepoint" for the character (the exceptions are those platforms where |
390 | the legacy encoding is some variant of EBCDIC rather than a super-set |
391 | of ASCII - see L<perlebcdic>). |
4411f3b6 |
392 | |
393 | Traditionaly computer data has been moved around in 8-bit chunks |
394 | often called "bytes". These chunks are also known as "octets" in |
395 | networking standards. Perl is widely used to manipulate data of |
396 | many types - not only strings of characters representing human or |
397 | computer languages but also "binary" data being the machines representation |
398 | of numbers, pixels in an image - or just about anything. |
399 | |
47bfe92f |
400 | When Perl is processing "binary data" the programmer wants Perl to process |
401 | "sequences of bytes". This is not a problem for Perl - as a byte has 256 |
402 | possible values it easily fits in Perl's much larger "logical character". |
4411f3b6 |
403 | |
404 | =head2 TERMINOLOGY |
405 | |
4ac9195f |
406 | =over 4 |
4411f3b6 |
407 | |
408 | =item * |
409 | |
410 | I<character>: a character in the range 0..(2**32-1) (or more). |
47bfe92f |
411 | (What Perl's strings are made of.) |
4411f3b6 |
412 | |
413 | =item * |
414 | |
415 | I<byte>: a character in the range 0..255 |
47bfe92f |
416 | (A special case of a Perl character.) |
4411f3b6 |
417 | |
418 | =item * |
419 | |
420 | I<octet>: 8 bits of data, with ordinal values 0..255 |
47bfe92f |
421 | (Term for bytes passed to or from a non-Perl context, e.g. disk file.) |
4411f3b6 |
422 | |
423 | =back |
424 | |
425 | The marker [INTERNAL] marks Internal Implementation Details, in |
426 | general meant only for those who think they know what they are doing, |
427 | and such details may change in future releases. |
428 | |
429 | =head1 ENCODINGS |
430 | |
431 | =head2 Characteristics of an Encoding |
432 | |
433 | An encoding has a "repertoire" of characters that it can represent, |
434 | and for each representable character there is at least one sequence of |
435 | octets that represents it. |
436 | |
437 | =head2 Types of Encodings |
438 | |
439 | Encodings can be divided into the following types: |
440 | |
441 | =over 4 |
442 | |
443 | =item * Fixed length 8-bit (or less) encodings. |
444 | |
445 | Each character is a single octet so may have a repertoire of up to |
446 | 256 characters. ASCII and iso-8859-* are typical examples. |
447 | |
448 | =item * Fixed length 16-bit encodings |
449 | |
450 | Each character is two octets so may have a repertoire of up to |
47bfe92f |
451 | 65 536 characters. Unicode's UCS-2 is an example. Also used for |
4411f3b6 |
452 | encodings for East Asian languages. |
453 | |
454 | =item * Fixed length 32-bit encodings. |
455 | |
456 | Not really very "encoded" encodings. The Unicode code points |
457 | are just represented as 4-octet integers. None the less because |
458 | different architectures use different representations of integers |
459 | (so called "endian") there at least two disctinct encodings. |
460 | |
461 | =item * Multi-byte encodings |
462 | |
463 | The number of octets needed to represent a character varies. |
464 | UTF-8 is a particularly complex but regular case of a multi-byte |
465 | encoding. Several East Asian countries use a multi-byte encoding |
466 | where 1-octet is used to cover western roman characters and Asian |
467 | characters get 2-octets. |
468 | (UTF-16 is strictly a multi-byte encoding taking either 2 or 4 octets |
469 | to represent a Unicode code point.) |
470 | |
471 | =item * "Escape" encodings. |
472 | |
473 | These encodings embed "escape sequences" into the octet sequence |
474 | which describe how the following octets are to be interpreted. |
475 | The iso-2022-* family is typical. Following the escape sequence |
476 | octets are encoded by an "embedded" encoding (which will be one |
477 | of the above types) until another escape sequence switches to |
478 | a different "embedded" encoding. |
479 | |
480 | These schemes are very flexible and can handle mixed languages but are |
47bfe92f |
481 | very complex to process (and have state). No escape encodings are |
482 | implemented for Perl yet. |
4411f3b6 |
483 | |
484 | =back |
485 | |
486 | =head2 Specifying Encodings |
487 | |
488 | Encodings can be specified to the API described below in two ways: |
489 | |
490 | =over 4 |
491 | |
492 | =item 1. By name |
493 | |
47bfe92f |
494 | Encoding names are strings with characters taken from a restricted |
495 | repertoire. See L</"Encoding Names">. |
4411f3b6 |
496 | |
497 | =item 2. As an object |
498 | |
499 | Encoding objects are returned by C<find_encoding($name)>. |
500 | |
501 | =back |
502 | |
503 | =head2 Encoding Names |
504 | |
505 | Encoding names are case insensitive. White space in names is ignored. |
47bfe92f |
506 | In addition an encoding may have aliases. Each encoding has one |
507 | "canonical" name. The "canonical" name is chosen from the names of |
508 | the encoding by picking the first in the following sequence: |
4411f3b6 |
509 | |
510 | =over 4 |
511 | |
512 | =item * The MIME name as defined in IETF RFC-XXXX. |
513 | |
514 | =item * The name in the IANA registry. |
515 | |
516 | =item * The name used by the the organization that defined it. |
517 | |
518 | =back |
519 | |
520 | Because of all the alias issues, and because in the general case |
521 | encodings have state C<Encode> uses the encoding object internally |
522 | once an operation is in progress. |
523 | |
4411f3b6 |
524 | =head1 PERL ENCODING API |
525 | |
526 | =head2 Generic Encoding Interface |
527 | |
528 | =over 4 |
529 | |
530 | =item * |
531 | |
532 | $bytes = encode(ENCODING, $string[, CHECK]) |
533 | |
47bfe92f |
534 | Encodes string from Perl's internal form into I<ENCODING> and returns |
535 | a sequence of octets. For CHECK see L</"Handling Malformed Data">. |
4411f3b6 |
536 | |
537 | =item * |
538 | |
539 | $string = decode(ENCODING, $bytes[, CHECK]) |
540 | |
47bfe92f |
541 | Decode sequence of octets assumed to be in I<ENCODING> into Perl's |
542 | internal form and returns the resulting string. For CHECK see |
543 | L</"Handling Malformed Data">. |
544 | |
545 | =item * |
546 | |
547 | from_to($string, FROM_ENCODING, TO_ENCODING[, CHECK]) |
548 | |
2b106fbe |
549 | Convert B<in-place> the data between two encodings. How did the data |
550 | in $string originally get to be in FROM_ENCODING? Either using |
e9692b5b |
551 | encode() or through PerlIO: See L</"Encoding and IO">. For CHECK |
2b106fbe |
552 | see L</"Handling Malformed Data">. |
553 | |
554 | For example to convert ISO 8859-1 data to UTF-8: |
555 | |
556 | from_to($data, "iso-8859-1", "utf-8"); |
557 | |
558 | and to convert it back: |
559 | |
560 | from_to($data, "utf-8", "iso-8859-1"); |
4411f3b6 |
561 | |
ab97ca19 |
562 | Note that because the conversion happens in place, the data to be |
563 | converted cannot be a string constant, it must be a scalar variable. |
564 | |
4411f3b6 |
565 | =back |
566 | |
567 | =head2 Handling Malformed Data |
568 | |
569 | If CHECK is not set, C<undef> is returned. If the data is supposed to |
47bfe92f |
570 | be UTF-8, an optional lexical warning (category utf8) is given. If |
571 | CHECK is true but not a code reference, dies. |
4411f3b6 |
572 | |
47bfe92f |
573 | It would desirable to have a way to indicate that transform should use |
574 | the encodings "replacement character" - no such mechanism is defined yet. |
4411f3b6 |
575 | |
576 | It is also planned to allow I<CHECK> to be a code reference. |
577 | |
47bfe92f |
578 | This is not yet implemented as there are design issues with what its |
579 | arguments should be and how it returns its results. |
4411f3b6 |
580 | |
581 | =over 4 |
582 | |
583 | =item Scheme 1 |
584 | |
585 | Passed remaining fragment of string being processed. |
586 | Modifies it in place to remove bytes/characters it can understand |
587 | and returns a string used to represent them. |
588 | e.g. |
589 | |
590 | sub fixup { |
591 | my $ch = substr($_[0],0,1,''); |
592 | return sprintf("\x{%02X}",ord($ch); |
593 | } |
594 | |
595 | This scheme is close to how underlying C code for Encode works, but gives |
596 | the fixup routine very little context. |
597 | |
598 | =item Scheme 2 |
599 | |
47bfe92f |
600 | Passed original string, and an index into it of the problem area, and |
601 | output string so far. Appends what it will to output string and |
602 | returns new index into original string. For example: |
4411f3b6 |
603 | |
604 | sub fixup { |
605 | # my ($s,$i,$d) = @_; |
606 | my $ch = substr($_[0],$_[1],1); |
607 | $_[2] .= sprintf("\x{%02X}",ord($ch); |
608 | return $_[1]+1; |
609 | } |
610 | |
47bfe92f |
611 | This scheme gives maximal control to the fixup routine but is more |
612 | complicated to code, and may need internals of Encode to be tweaked to |
613 | keep original string intact. |
4411f3b6 |
614 | |
615 | =item Other Schemes |
616 | |
617 | Hybrids of above. |
618 | |
619 | Multiple return values rather than in-place modifications. |
620 | |
621 | Index into the string could be pos($str) allowing s/\G...//. |
622 | |
623 | =back |
624 | |
625 | =head2 UTF-8 / utf8 |
626 | |
627 | The Unicode consortium defines the UTF-8 standard as a way of encoding |
47bfe92f |
628 | the entire Unicode repertiore as sequences of octets. This encoding is |
629 | expected to become very widespread. Perl can use this form internaly |
630 | to represent strings, so conversions to and from this form are |
631 | particularly efficient (as octets in memory do not have to change, |
632 | just the meta-data that tells Perl how to treat them). |
4411f3b6 |
633 | |
634 | =over 4 |
635 | |
636 | =item * |
637 | |
638 | $bytes = encode_utf8($string); |
639 | |
47bfe92f |
640 | The characters that comprise string are encoded in Perl's superset of UTF-8 |
4411f3b6 |
641 | and the resulting octets returned as a sequence of bytes. All possible |
642 | characters have a UTF-8 representation so this function cannot fail. |
643 | |
644 | =item * |
645 | |
646 | $string = decode_utf8($bytes [,CHECK]); |
647 | |
47bfe92f |
648 | The sequence of octets represented by $bytes is decoded from UTF-8 |
649 | into a sequence of logical characters. Not all sequences of octets |
650 | form valid UTF-8 encodings, so it is possible for this call to fail. |
651 | For CHECK see L</"Handling Malformed Data">. |
4411f3b6 |
652 | |
653 | =back |
654 | |
655 | =head2 Other Encodings of Unicode |
656 | |
47bfe92f |
657 | UTF-16 is similar to UCS-2, 16 bit or 2-byte chunks. UCS-2 can only |
658 | represent 0..0xFFFF, while UTF-16 has a "surrogate pair" scheme which |
659 | allows it to cover the whole Unicode range. |
4411f3b6 |
660 | |
8040349a |
661 | Encode implements big-endian UCS-2 aliased to "iso-10646-1" as that |
47bfe92f |
662 | happens to be the name used by that representation when used with X11 |
663 | fonts. |
4411f3b6 |
664 | |
665 | UTF-32 or UCS-4 is 32-bit or 4-byte chunks. Perl's logical characters |
666 | can be considered as being in this form without encoding. An encoding |
47bfe92f |
667 | to transfer strings in this form (e.g. to write them to a file) would |
668 | need to |
4411f3b6 |
669 | |
670 | pack('L',map(chr($_),split(//,$string))); # native |
671 | or |
672 | pack('V',map(chr($_),split(//,$string))); # little-endian |
673 | or |
674 | pack('N',map(chr($_),split(//,$string))); # big-endian |
675 | |
676 | depending on the endian required. |
677 | |
51ef4e11 |
678 | No UTF-32 encodings are implemented yet. |
4411f3b6 |
679 | |
47bfe92f |
680 | Both UCS-2 and UCS-4 style encodings can have "byte order marks" by |
681 | representing the code point 0xFFFE as the very first thing in a file. |
4411f3b6 |
682 | |
51ef4e11 |
683 | =head2 Listing available encodings |
684 | |
685 | use Encode qw(encodings); |
686 | @list = encodings(); |
687 | |
688 | Returns a list of the canonical names of the available encodings. |
689 | |
690 | =head2 Defining Aliases |
691 | |
692 | use Encode qw(define_alias); |
693 | define_alias( newName => ENCODING); |
694 | |
47bfe92f |
695 | Allows newName to be used as am alias for ENCODING. ENCODING may be |
696 | either the name of an encoding or and encoding object (as above). |
51ef4e11 |
697 | |
698 | Currently I<newName> can be specified in the following ways: |
699 | |
700 | =over 4 |
701 | |
702 | =item As a simple string. |
703 | |
704 | =item As a qr// compiled regular expression, e.g.: |
705 | |
706 | define_alias( qr/^iso8859-(\d+)$/i => '"iso-8859-$1"' ); |
707 | |
47bfe92f |
708 | In this case if I<ENCODING> is not a reference it is C<eval>-ed to |
709 | allow C<$1> etc. to be subsituted. The example is one way to names as |
710 | used in X11 font names to alias the MIME names for the iso-8859-* |
711 | family. |
51ef4e11 |
712 | |
713 | =item As a code reference, e.g.: |
714 | |
715 | define_alias( sub { return /^iso8859-(\d+)$/i ? "iso-8859-$1" : undef } , ''); |
716 | |
717 | In this case C<$_> will be set to the name that is being looked up and |
47bfe92f |
718 | I<ENCODING> is passed to the sub as its first argument. The example |
719 | is another way to names as used in X11 font names to alias the MIME |
720 | names for the iso-8859-* family. |
51ef4e11 |
721 | |
722 | =back |
723 | |
724 | =head2 Defining Encodings |
725 | |
e9692b5b |
726 | use Encode qw(define_alias); |
727 | define_encoding( $object, 'canonicalName' [,alias...]); |
51ef4e11 |
728 | |
47bfe92f |
729 | Causes I<canonicalName> to be associated with I<$object>. The object |
730 | should provide the interface described in L</"IMPLEMENTATION CLASSES"> |
731 | below. If more than two arguments are provided then additional |
732 | arguments are taken as aliases for I<$object> as for C<define_alias>. |
51ef4e11 |
733 | |
4411f3b6 |
734 | =head1 Encoding and IO |
735 | |
736 | It is very common to want to do encoding transformations when |
737 | reading or writing files, network connections, pipes etc. |
47bfe92f |
738 | If Perl is configured to use the new 'perlio' IO system then |
4411f3b6 |
739 | C<Encode> provides a "layer" (See L<perliol>) which can transform |
740 | data as it is read or written. |
741 | |
8e86646e |
742 | Here is how the blind poet would modernise the encoding: |
743 | |
42234700 |
744 | use Encode; |
8e86646e |
745 | open(my $iliad,'<:encoding(iso-8859-7)','iliad.greek'); |
746 | open(my $utf8,'>:utf8','iliad.utf8'); |
747 | my @epic = <$iliad>; |
748 | print $utf8 @epic; |
749 | close($utf8); |
750 | close($illiad); |
4411f3b6 |
751 | |
752 | In addition the new IO system can also be configured to read/write |
753 | UTF-8 encoded characters (as noted above this is efficient): |
754 | |
e9692b5b |
755 | open(my $fh,'>:utf8','anything'); |
756 | print $fh "Any \x{0021} string \N{SMILEY FACE}\n"; |
4411f3b6 |
757 | |
758 | Either of the above forms of "layer" specifications can be made the default |
759 | for a lexical scope with the C<use open ...> pragma. See L<open>. |
760 | |
761 | Once a handle is open is layers can be altered using C<binmode>. |
762 | |
47bfe92f |
763 | Without any such configuration, or if Perl itself is built using |
4411f3b6 |
764 | system's own IO, then write operations assume that file handle accepts |
765 | only I<bytes> and will C<die> if a character larger than 255 is |
766 | written to the handle. When reading, each octet from the handle |
767 | becomes a byte-in-a-character. Note that this default is the same |
47bfe92f |
768 | behaviour as bytes-only languages (including Perl before v5.6) would |
769 | have, and is sufficient to handle native 8-bit encodings |
770 | e.g. iso-8859-1, EBCDIC etc. and any legacy mechanisms for handling |
771 | other encodings and binary data. |
772 | |
773 | In other cases it is the programs responsibility to transform |
774 | characters into bytes using the API above before doing writes, and to |
775 | transform the bytes read from a handle into characters before doing |
776 | "character operations" (e.g. C<lc>, C</\W+/>, ...). |
777 | |
47bfe92f |
778 | You can also use PerlIO to convert larger amounts of data you don't |
779 | want to bring into memory. For example to convert between ISO 8859-1 |
780 | (Latin 1) and UTF-8 (or UTF-EBCDIC in EBCDIC machines): |
781 | |
e9692b5b |
782 | open(F, "<:encoding(iso-8859-1)", "data.txt") or die $!; |
783 | open(G, ">:utf8", "data.utf") or die $!; |
784 | while (<F>) { print G } |
785 | |
786 | # Could also do "print G <F>" but that would pull |
787 | # the whole file into memory just to write it out again. |
788 | |
789 | More examples: |
47bfe92f |
790 | |
e9692b5b |
791 | open(my $f, "<:encoding(cp1252)") |
792 | open(my $g, ">:encoding(iso-8859-2)") |
793 | open(my $h, ">:encoding(latin9)") # iso-8859-15 |
47bfe92f |
794 | |
795 | See L<PerlIO> for more information. |
4411f3b6 |
796 | |
797 | =head1 Encoding How to ... |
798 | |
799 | To do: |
800 | |
801 | =over 4 |
802 | |
803 | =item * IO with mixed content (faking iso-2020-*) |
804 | |
805 | =item * MIME's Content-Length: |
806 | |
807 | =item * UTF-8 strings in binary data. |
808 | |
47bfe92f |
809 | =item * Perl/Encode wrappers on non-Unicode XS modules. |
4411f3b6 |
810 | |
811 | =back |
812 | |
813 | =head1 Messing with Perl's Internals |
814 | |
47bfe92f |
815 | The following API uses parts of Perl's internals in the current |
816 | implementation. As such they are efficient, but may change. |
4411f3b6 |
817 | |
818 | =over 4 |
819 | |
4411f3b6 |
820 | =item * is_utf8(STRING [, CHECK]) |
821 | |
822 | [INTERNAL] Test whether the UTF-8 flag is turned on in the STRING. |
47bfe92f |
823 | If CHECK is true, also checks the data in STRING for being well-formed |
824 | UTF-8. Returns true if successful, false otherwise. |
4411f3b6 |
825 | |
826 | =item * valid_utf8(STRING) |
827 | |
47bfe92f |
828 | [INTERNAL] Test whether STRING is in a consistent state. Will return |
829 | true if string is held as bytes, or is well-formed UTF-8 and has the |
830 | UTF-8 flag on. Main reason for this routine is to allow Perl's |
831 | testsuite to check that operations have left strings in a consistent |
832 | state. |
4411f3b6 |
833 | |
834 | =item * |
835 | |
836 | _utf8_on(STRING) |
837 | |
838 | [INTERNAL] Turn on the UTF-8 flag in STRING. The data in STRING is |
839 | B<not> checked for being well-formed UTF-8. Do not use unless you |
840 | B<know> that the STRING is well-formed UTF-8. Returns the previous |
841 | state of the UTF-8 flag (so please don't test the return value as |
842 | I<not> success or failure), or C<undef> if STRING is not a string. |
843 | |
844 | =item * |
845 | |
846 | _utf8_off(STRING) |
847 | |
848 | [INTERNAL] Turn off the UTF-8 flag in STRING. Do not use frivolously. |
849 | Returns the previous state of the UTF-8 flag (so please don't test the |
850 | return value as I<not> success or failure), or C<undef> if STRING is |
851 | not a string. |
852 | |
853 | =back |
854 | |
4edaa979 |
855 | =head1 IMPLEMENTATION CLASSES |
856 | |
857 | As mentioned above encodings are (in the current implementation at least) |
858 | defined by objects. The mapping of encoding name to object is via the |
51ef4e11 |
859 | C<%encodings> hash. |
4edaa979 |
860 | |
861 | The values of the hash can currently be either strings or objects. |
862 | The string form may go away in the future. The string form occurs |
863 | when C<encodings()> has scanned C<@INC> for loadable encodings but has |
864 | not actually loaded the encoding in question. This is because the |
47bfe92f |
865 | current "loading" process is all Perl and a bit slow. |
4edaa979 |
866 | |
47bfe92f |
867 | Once an encoding is loaded then value of the hash is object which |
868 | implements the encoding. The object should provide the following |
869 | interface: |
4edaa979 |
870 | |
871 | =over 4 |
872 | |
873 | =item -E<gt>name |
874 | |
875 | Should return the string representing the canonical name of the encoding. |
876 | |
877 | =item -E<gt>new_sequence |
878 | |
47bfe92f |
879 | This is a placeholder for encodings with state. It should return an |
880 | object which implements this interface, all current implementations |
881 | return the original object. |
4edaa979 |
882 | |
883 | =item -E<gt>encode($string,$check) |
884 | |
47bfe92f |
885 | Should return the octet sequence representing I<$string>. If I<$check> |
886 | is true it should modify I<$string> in place to remove the converted |
887 | part (i.e. the whole string unless there is an error). If an error |
888 | occurs it should return the octet sequence for the fragment of string |
889 | that has been converted, and modify $string in-place to remove the |
890 | converted part leaving it starting with the problem fragment. |
4edaa979 |
891 | |
47bfe92f |
892 | If check is is false then C<encode> should make a "best effort" to |
893 | convert the string - for example by using a replacement character. |
4edaa979 |
894 | |
895 | =item -E<gt>decode($octets,$check) |
896 | |
47bfe92f |
897 | Should return the string that I<$octets> represents. If I<$check> is |
898 | true it should modify I<$octets> in place to remove the converted part |
899 | (i.e. the whole sequence unless there is an error). If an error |
900 | occurs it should return the fragment of string that has been |
901 | converted, and modify $octets in-place to remove the converted part |
4edaa979 |
902 | leaving it starting with the problem fragment. |
903 | |
47bfe92f |
904 | If check is is false then C<decode> should make a "best effort" to |
905 | convert the string - for example by using Unicode's "\x{FFFD}" as a |
906 | replacement character. |
4edaa979 |
907 | |
908 | =back |
909 | |
47bfe92f |
910 | It should be noted that the check behaviour is different from the |
911 | outer public API. The logic is that the "unchecked" case is useful |
912 | when encoding is part of a stream which may be reporting errors |
913 | (e.g. STDERR). In such cases it is desirable to get everything |
914 | through somehow without causing additional errors which obscure the |
915 | original one. Also the encoding is best placed to know what the |
916 | correct replacement character is, so if that is the desired behaviour |
917 | then letting low level code do it is the most efficient. |
918 | |
919 | In contrast if check is true, the scheme above allows the encoding to |
920 | do as much as it can and tell layer above how much that was. What is |
921 | lacking at present is a mechanism to report what went wrong. The most |
922 | likely interface will be an additional method call to the object, or |
923 | perhaps (to avoid forcing per-stream objects on otherwise stateless |
924 | encodings) and additional parameter. |
925 | |
926 | It is also highly desirable that encoding classes inherit from |
927 | C<Encode::Encoding> as a base class. This allows that class to define |
928 | additional behaviour for all encoding objects. For example built in |
929 | Unicode, UCS-2 and UTF-8 classes use : |
51ef4e11 |
930 | |
931 | package Encode::MyEncoding; |
932 | use base qw(Encode::Encoding); |
933 | |
934 | __PACKAGE__->Define(qw(myCanonical myAlias)); |
935 | |
47bfe92f |
936 | To create an object with bless {Name => ...},$class, and call |
937 | define_encoding. They inherit their C<name> method from |
938 | C<Encode::Encoding>. |
4edaa979 |
939 | |
940 | =head2 Compiled Encodings |
941 | |
47bfe92f |
942 | F<Encode.xs> provides a class C<Encode::XS> which provides the |
943 | interface described above. It calls a generic octet-sequence to |
944 | octet-sequence "engine" that is driven by tables (defined in |
945 | F<encengine.c>). The same engine is used for both encode and |
946 | decode. C<Encode:XS>'s C<encode> forces Perl's characters to their |
947 | UTF-8 form and then treats them as just another multibyte |
948 | encoding. C<Encode:XS>'s C<decode> transforms the sequence and then |
949 | turns the UTF-8-ness flag as that is the form that the tables are |
950 | defined to produce. For details of the engine see the comments in |
951 | F<encengine.c>. |
952 | |
953 | The tables are produced by the Perl script F<compile> (the name needs |
954 | to change so we can eventually install it somewhere). F<compile> can |
955 | currently read two formats: |
4edaa979 |
956 | |
957 | =over 4 |
958 | |
959 | =item *.enc |
960 | |
47bfe92f |
961 | This is a coined format used by Tcl. It is documented in |
962 | Encode/EncodeFormat.pod. |
4edaa979 |
963 | |
964 | =item *.ucm |
965 | |
966 | This is the semi-standard format used by IBM's ICU package. |
967 | |
968 | =back |
969 | |
970 | F<compile> can write the following forms: |
971 | |
972 | =over 4 |
973 | |
974 | =item *.ucm |
975 | |
976 | See above - the F<Encode/*.ucm> files provided with the distribution have |
977 | been created from the original Tcl .enc files using this approach. |
978 | |
979 | =item *.c |
980 | |
981 | Produces tables as C data structures - this is used to build in encodings |
982 | into F<Encode.so>/F<Encode.dll>. |
983 | |
984 | =item *.xs |
985 | |
47bfe92f |
986 | In theory this allows encodings to be stand-alone loadable Perl |
987 | extensions. The process has not yet been tested. The plan is to use |
988 | this approach for large East Asian encodings. |
4edaa979 |
989 | |
990 | =back |
991 | |
47bfe92f |
992 | The set of encodings built-in to F<Encode.so>/F<Encode.dll> is |
993 | determined by F<Makefile.PL>. The current set is as follows: |
4edaa979 |
994 | |
995 | =over 4 |
996 | |
997 | =item ascii and iso-8859-* |
998 | |
999 | That is all the common 8-bit "western" encodings. |
1000 | |
1001 | =item IBM-1047 and two other variants of EBCDIC. |
1002 | |
47bfe92f |
1003 | These are the same variants that are supported by EBCDIC Perl as |
1004 | "native" encodings. They are included to prove "reversibility" of |
1005 | some constructs in EBCDIC Perl. |
4edaa979 |
1006 | |
1007 | =item symbol and dingbats as used by Tk on X11. |
1008 | |
47bfe92f |
1009 | (The reason Encode got started was to support Perl/Tk.) |
4edaa979 |
1010 | |
1011 | =back |
1012 | |
47bfe92f |
1013 | That set is rather ad hoc and has been driven by the needs of the |
1014 | tests rather than the needs of typical applications. It is likely |
1015 | to be rationalized. |
4edaa979 |
1016 | |
4411f3b6 |
1017 | =head1 SEE ALSO |
1018 | |
47bfe92f |
1019 | L<perlunicode>, L<perlebcdic>, L<perlfunc/open>, L<PerlIO> |
4411f3b6 |
1020 | |
1021 | =cut |
1022 | |