8 our @ISA = qw(Exporter DynaLoader);
10 # Public, encouraged API is exported by default
36 # Documentation moved after __END__ for speed - NI-S
40 # Make a %encoding package variable to allow a certain amount of cheating
42 my @alias; # ordered matching list
43 my %alias; # cached known aliases
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 );
65 sort { $a->[1] cmp $b->[1] }
67 grep { $_ ne 'Internal' }
75 # print "# findAlias $_\n";
76 unless (exists $alias{$_})
78 for (my $i=0; $i < @alias; $i += 2)
80 my $alias = $alias[$i];
81 my $val = $alias[$i+1];
83 if (ref($alias) eq 'Regexp' && $_ =~ $alias)
87 elsif (ref($alias) eq 'CODE')
89 $new = &{$alias}($val)
91 elsif (lc($_) eq lc($alias))
97 next if $new eq $_; # avoid (direct) recursion on bugs
98 my $enc = (ref($new)) ? $new : find_encoding($new);
114 my ($alias,$name) = splice(@_,0,2);
115 push(@alias, $alias => $name);
119 # Allow variants of iso-8859-1 etc.
120 define_alias( qr/^iso[-_]?(\d+)[-_](\d+)$/i => '"iso-$1-$2"' );
122 # At least HP-UX has these.
123 define_alias( qr/^iso8859(\d+)$/i => '"iso-8859-$1"' );
126 define_alias( qr/^(?:hp-)?(arabic|greek|hebrew|kana|roman|thai|turkish)8$/i => '"${1}8"' );
128 # The Official name of ASCII.
129 define_alias( qr/^ANSI[-_]?X3\.4[-_]?1968$/i => '"ascii"' );
131 # This is a font issue, not an encoding issue.
132 # (The currency symbol of the Latin 1 upper half
133 # has been redefined as the euro symbol.)
134 define_alias( qr/^(.+)\@euro$/i => '"$1"' );
136 # Allow latin-1 style names as well
137 define_alias( qr/^(?:iso[-_]?)?latin[-_]?(\d+)$/i => '"iso-8859-$latin2iso_num[$1]"' );
139 # Allow winlatin1 style names as well
140 define_alias( qr/^win(latin[12]|cyrillic|baltic|greek|turkish|hebrew|arabic|baltic|vietnamese)$/i => '"cp$winlatin2cp{\u$1}"' );
142 # Common names for non-latin prefered MIME names
143 define_alias( 'ascii' => 'US-ascii',
144 'cyrillic' => 'iso-8859-5',
145 'arabic' => 'iso-8859-6',
146 'greek' => 'iso-8859-7',
147 'hebrew' => 'iso-8859-8',
148 'thai' => 'iso-8859-11',
149 'tis620' => 'iso-8859-11',
152 # At least AIX has IBM-NNN (surprisingly...) instead of cpNNN.
153 # And Microsoft has their own naming (again, surprisingly).
154 define_alias( qr/^(?:ibm|ms)[-_]?(\d\d\d\d?)$/i => '"cp$1"');
156 # Sometimes seen with a leading zero.
157 define_alias( qr/^cp037$/i => '"cp37"');
160 define_alias( qr/^macRomanian$/i => '"macRumanian"');
162 # Standardize on the dashed versions.
163 define_alias( qr/^utf8$/i => 'utf-8' );
164 define_alias( qr/^koi8r$/i => 'koi8-r' );
165 define_alias( qr/^koi8u$/i => 'koi8-u' );
167 # Seen in some Linuxes.
168 define_alias( qr/^ujis$/i => 'euc-jp' );
170 # CP936 doesn't have vendor-addon for GBK, so they're identical.
171 define_alias( qr/^gbk$/i => '"cp936"');
173 # TODO: HP-UX '8' encodings arabic8 greek8 hebrew8 kana8 thai8 turkish8
174 # TODO: HP-UX '15' encodings japanese15 korean15 roi15
175 # TODO: Cyrillic encoding ISO-IR-111 (useful?)
176 # TODO: Chinese encodings HZ
177 # TODO: Armenian encoding ARMSCII-8
178 # TODO: Hebrew encoding ISO-8859-8-1
179 # TODO: Thai encoding TCVN
180 # TODO: Korean encoding Johab
181 # TODO: Vietnamese encodings VPS
182 # TODO: Japanese encoding JIS (not the same as SJIS)
183 # TODO: Mac Asian+African encodings: Arabic Armenian Bengali Burmese
184 # ChineseSimp ChineseTrad Devanagari Ethiopic ExtArabic
185 # Farsi Georgian Gujarati Gurmukhi Hebrew Japanese
186 # Kannada Khmer Korean Laotian Malayalam Mongolian
187 # Oriya Sinhalese Symbol Tamil Telugu Tibetan Vietnamese
189 # Map white space and _ to '-'
190 define_alias( qr/^(\S+)[\s_]+(.*)$/i => '"$1-$2"' );
196 $encoding{$name} = $obj;
198 define_alias($lc => $obj) unless $lc eq $name;
202 define_alias($alias,$obj);
209 my ($class,$name) = @_;
211 if (ref($name) && $name->can('new_sequence'))
216 if (exists $encoding{$name})
218 return $encoding{$name};
220 if (exists $encoding{$lc})
222 return $encoding{$lc};
225 my $oc = $class->findAlias($name);
226 return $oc if defined $oc;
227 return $class->findAlias($lc) if $lc ne $name;
235 return __PACKAGE__->getEncoding($name);
240 my ($name,$string,$check) = @_;
241 my $enc = find_encoding($name);
242 croak("Unknown encoding '$name'") unless defined $enc;
243 my $octets = $enc->encode($string,$check);
244 return undef if ($check && length($string));
250 my ($name,$octets,$check) = @_;
251 my $enc = find_encoding($name);
252 croak("Unknown encoding '$name'") unless defined $enc;
253 my $string = $enc->decode($octets,$check);
254 $_[1] = $octets if $check;
260 my ($string,$from,$to,$check) = @_;
261 my $f = find_encoding($from);
262 croak("Unknown encoding '$from'") unless defined $f;
263 my $t = find_encoding($to);
264 croak("Unknown encoding '$to'") unless defined $t;
265 my $uni = $f->decode($string,$check);
266 return undef if ($check && length($string));
267 $string = $t->encode($uni,$check);
268 return undef if ($check && length($uni));
269 return length($_[0] = $string);
282 return undef unless utf8::decode($str);
286 require Encode::Encoding;
288 require Encode::Internal;
289 require Encode::Unicode;
290 require Encode::utf8;
291 require Encode::iso10646_1;
292 require Encode::ucs2_le;
300 Encode - character encodings
306 use Encode::TW; # for Taiwan-based Chinese encodings
307 use Encode::CN; # for China-based Chinese encodings
308 use Encode::JP; # for Japanese encodings
309 use Encode::KR; # for Korean encodings
313 The C<Encode> module provides the interfaces between Perl's strings
314 and the rest of the system. Perl strings are sequences of B<characters>.
316 The repertoire of characters that Perl can represent is at least that
317 defined by the Unicode Consortium. On most platforms the ordinal
318 values of the characters (as returned by C<ord(ch)>) is the "Unicode
319 codepoint" for the character (the exceptions are those platforms where
320 the legacy encoding is some variant of EBCDIC rather than a super-set
321 of ASCII - see L<perlebcdic>).
323 Traditionaly computer data has been moved around in 8-bit chunks
324 often called "bytes". These chunks are also known as "octets" in
325 networking standards. Perl is widely used to manipulate data of
326 many types - not only strings of characters representing human or
327 computer languages but also "binary" data being the machines representation
328 of numbers, pixels in an image - or just about anything.
330 When Perl is processing "binary data" the programmer wants Perl to process
331 "sequences of bytes". This is not a problem for Perl - as a byte has 256
332 possible values it easily fits in Perl's much larger "logical character".
334 Due to size concerns, before using B<CJK> (Chinese, Japanese & Korean)
335 encodings, you have to C<use> the corresponding
336 B<Encode::>(B<TW>|B<CN>|B<JP>|B<KR>) modules first.
344 I<character>: a character in the range 0..(2**32-1) (or more).
345 (What Perl's strings are made of.)
349 I<byte>: a character in the range 0..255
350 (A special case of a Perl character.)
354 I<octet>: 8 bits of data, with ordinal values 0..255
355 (Term for bytes passed to or from a non-Perl context, e.g. disk file.)
359 The marker [INTERNAL] marks Internal Implementation Details, in
360 general meant only for those who think they know what they are doing,
361 and such details may change in future releases.
365 =head2 Characteristics of an Encoding
367 An encoding has a "repertoire" of characters that it can represent,
368 and for each representable character there is at least one sequence of
369 octets that represents it.
371 =head2 Types of Encodings
373 Encodings can be divided into the following types:
377 =item * Fixed length 8-bit (or less) encodings.
379 Each character is a single octet so may have a repertoire of up to
380 256 characters. ASCII and iso-8859-* are typical examples.
382 =item * Fixed length 16-bit encodings
384 Each character is two octets so may have a repertoire of up to
385 65 536 characters. Unicode's UCS-2 is an example. Also used for
386 encodings for East Asian languages.
388 =item * Fixed length 32-bit encodings.
390 Not really very "encoded" encodings. The Unicode code points
391 are just represented as 4-octet integers. None the less because
392 different architectures use different representations of integers
393 (so called "endian") there at least two disctinct encodings.
395 =item * Multi-byte encodings
397 The number of octets needed to represent a character varies.
398 UTF-8 is a particularly complex but regular case of a multi-byte
399 encoding. Several East Asian countries use a multi-byte encoding
400 where 1-octet is used to cover western roman characters and Asian
401 characters get 2-octets.
402 (UTF-16 is strictly a multi-byte encoding taking either 2 or 4 octets
403 to represent a Unicode code point.)
405 =item * "Escape" encodings.
407 These encodings embed "escape sequences" into the octet sequence
408 which describe how the following octets are to be interpreted.
409 The iso-2022-* family is typical. Following the escape sequence
410 octets are encoded by an "embedded" encoding (which will be one
411 of the above types) until another escape sequence switches to
412 a different "embedded" encoding.
414 These schemes are very flexible and can handle mixed languages but are
415 very complex to process (and have state). No escape encodings are
416 implemented for Perl yet.
420 =head2 Specifying Encodings
422 Encodings can be specified to the API described below in two ways:
428 Encoding names are strings with characters taken from a restricted
429 repertoire. See L</"Encoding Names">.
431 =item 2. As an object
433 Encoding objects are returned by C<find_encoding($name)>.
437 =head2 Encoding Names
439 Encoding names are case insensitive. White space in names is ignored.
440 In addition an encoding may have aliases. Each encoding has one
441 "canonical" name. The "canonical" name is chosen from the names of
442 the encoding by picking the first in the following sequence:
446 =item * The MIME name as defined in IETF RFCs.
448 =item * The name in the IANA registry.
450 =item * The name used by the organization that defined it.
454 Because of all the alias issues, and because in the general case
455 encodings have state C<Encode> uses the encoding object internally
456 once an operation is in progress.
458 As of Perl 5.8.0, at least the following encodings are recognized
459 (the => marks aliases):
473 The ISO 8859 and KOI:
475 ISO 8859-1 ISO 8859-6 ISO 8859-11 KOI8-F
476 ISO 8859-2 ISO 8859-7 (12 doesn't exist) KOI8-R
477 ISO 8859-3 ISO 8859-8 ISO 8859-13 KOI8-U
478 ISO 8859-4 ISO 8859-9 ISO 8859-14
479 ISO 8859-5 ISO 8859-10 ISO 8859-15
482 Latin1 => 8859-1 Latin6 => 8859-10
483 Latin2 => 8859-2 Latin7 => 8859-13
484 Latin3 => 8859-3 Latin8 => 8859-14
485 Latin4 => 8859-4 Latin9 => 8859-15
486 Latin5 => 8859-9 Latin10 => 8859-16
495 The CJKV: Chinese, Japanese, Korean, Vietnamese:
497 ISO 2022 ISO 2022 JP-1 JIS 0201 GB 1988 Big5 EUC-CN
498 ISO 2022 CN ISO 2022 JP-2 JIS 0208 GB 2312 HZ EUC-JP
499 ISO 2022 JP ISO 2022 KR JIS 0210 GB 12345 CNS 11643 EUC-JP-0212
500 Shift-JIS GBK Big5-HKSCS EUC-KR
503 (Due to size concerns, additional Chinese encodings including C<GB 18030>,
504 C<EUC-TW> and C<BIG5PLUS> are distributed separately on CPAN, under the name
505 L<Encode::HanExtra>.)
509 CP37 CP852 CP861 CP866 CP949 CP1251 CP1256
510 CP424 CP855 CP862 CP869 CP950 CP1252 CP1257
511 CP737 CP856 CP863 CP874 CP1006 CP1253 CP1258
512 CP775 CP857 CP864 CP932 CP1047 CP1254
513 CP850 CP860 CP865 CP936 CP1250 CP1255
517 WinCyrillic => CP1251
519 WinTurkiskh => CP1254
523 WinVietnamese => CP1258
525 (All the CPI<NNN...> are available also as IBMI<NNN...>.)
529 MacCentralEuropean MacJapanese
531 MacCyrillic MacRomanian
534 MacIcelandic MacTurkish
545 =head1 PERL ENCODING API
547 =head2 Generic Encoding Interface
553 $bytes = encode(ENCODING, $string[, CHECK])
555 Encodes string from Perl's internal form into I<ENCODING> and returns
556 a sequence of octets. For CHECK see L</"Handling Malformed Data">.
558 For example to convert (internally UTF-8 encoded) Unicode data
561 $octets = encode("utf8", $unicode);
565 $string = decode(ENCODING, $bytes[, CHECK])
567 Decode sequence of octets assumed to be in I<ENCODING> into Perl's
568 internal form and returns the resulting string. For CHECK see
569 L</"Handling Malformed Data">.
571 For example to convert ISO 8859-1 data to UTF-8:
573 $utf8 = decode("latin1", $latin1);
577 from_to($string, FROM_ENCODING, TO_ENCODING[, CHECK])
579 Convert B<in-place> the data between two encodings. How did the data
580 in $string originally get to be in FROM_ENCODING? Either using
581 encode() or through PerlIO: See L</"Encoding and IO">. For CHECK
582 see L</"Handling Malformed Data">.
584 For example to convert ISO 8859-1 data to UTF-8:
586 from_to($data, "iso-8859-1", "utf-8");
588 and to convert it back:
590 from_to($data, "utf-8", "iso-8859-1");
592 Note that because the conversion happens in place, the data to be
593 converted cannot be a string constant, it must be a scalar variable.
597 =head2 Handling Malformed Data
599 If CHECK is not set, C<undef> is returned. If the data is supposed to
600 be UTF-8, an optional lexical warning (category utf8) is given. If
601 CHECK is true but not a code reference, dies.
603 It would desirable to have a way to indicate that transform should use
604 the encodings "replacement character" - no such mechanism is defined yet.
606 It is also planned to allow I<CHECK> to be a code reference.
608 This is not yet implemented as there are design issues with what its
609 arguments should be and how it returns its results.
615 Passed remaining fragment of string being processed.
616 Modifies it in place to remove bytes/characters it can understand
617 and returns a string used to represent them.
621 my $ch = substr($_[0],0,1,'');
622 return sprintf("\x{%02X}",ord($ch);
625 This scheme is close to how underlying C code for Encode works, but gives
626 the fixup routine very little context.
630 Passed original string, and an index into it of the problem area, and
631 output string so far. Appends what it will to output string and
632 returns new index into original string. For example:
635 # my ($s,$i,$d) = @_;
636 my $ch = substr($_[0],$_[1],1);
637 $_[2] .= sprintf("\x{%02X}",ord($ch);
641 This scheme gives maximal control to the fixup routine but is more
642 complicated to code, and may need internals of Encode to be tweaked to
643 keep original string intact.
649 Multiple return values rather than in-place modifications.
651 Index into the string could be pos($str) allowing s/\G...//.
657 The Unicode consortium defines the UTF-8 standard as a way of encoding
658 the entire Unicode repertiore as sequences of octets. This encoding is
659 expected to become very widespread. Perl can use this form internaly
660 to represent strings, so conversions to and from this form are
661 particularly efficient (as octets in memory do not have to change,
662 just the meta-data that tells Perl how to treat them).
668 $bytes = encode_utf8($string);
670 The characters that comprise string are encoded in Perl's superset of UTF-8
671 and the resulting octets returned as a sequence of bytes. All possible
672 characters have a UTF-8 representation so this function cannot fail.
676 $string = decode_utf8($bytes [,CHECK]);
678 The sequence of octets represented by $bytes is decoded from UTF-8
679 into a sequence of logical characters. Not all sequences of octets
680 form valid UTF-8 encodings, so it is possible for this call to fail.
681 For CHECK see L</"Handling Malformed Data">.
685 =head2 Other Encodings of Unicode
687 UTF-16 is similar to UCS-2, 16 bit or 2-byte chunks. UCS-2 can only
688 represent 0..0xFFFF, while UTF-16 has a I<surrogate pair> scheme which
689 allows it to cover the whole Unicode range.
691 Surrogates are code points set aside to encode the 0x01000..0x10FFFF
692 range of Unicode code points in pairs of 16-bit units. The I<high
693 surrogates> are the range 0xD800..0xDBFF, and the I<low surrogates>
694 are the range 0xDC00..0xDFFFF. The surrogate encoding is
696 $hi = ($uni - 0x10000) / 0x400 + 0xD800;
697 $lo = ($uni - 0x10000) % 0x400 + 0xDC00;
701 $uni = 0x10000 + ($hi - 0xD8000) * 0x400 + ($lo - 0xDC00);
703 Encode implements big-endian UCS-2 aliased to "iso-10646-1" as that
704 happens to be the name used by that representation when used with X11
707 UTF-32 or UCS-4 is 32-bit or 4-byte chunks. Perl's logical characters
708 can be considered as being in this form without encoding. An encoding
709 to transfer strings in this form (e.g. to write them to a file) would
712 pack('L*', unpack('U*', $string)); # native
714 pack('V*', unpack('U*', $string)); # little-endian
716 pack('N*', unpack('U*', $string)); # big-endian
718 depending on the endianness required.
720 No UTF-32 encodings are implemented yet.
722 Both UCS-2 and UCS-4 style encodings can have "byte order marks" by
723 representing the code point 0xFFFE as the very first thing in a file.
725 =head2 Listing available encodings
727 use Encode qw(encodings);
730 Returns a list of the canonical names of the available encodings.
732 =head2 Defining Aliases
734 use Encode qw(define_alias);
735 define_alias( newName => ENCODING);
737 Allows newName to be used as am alias for ENCODING. ENCODING may be
738 either the name of an encoding or and encoding object (as above).
740 Currently I<newName> can be specified in the following ways:
744 =item As a simple string.
746 =item As a qr// compiled regular expression, e.g.:
748 define_alias( qr/^iso8859-(\d+)$/i => '"iso-8859-$1"' );
750 In this case if I<ENCODING> is not a reference it is C<eval>-ed to
751 allow C<$1> etc. to be subsituted. The example is one way to names as
752 used in X11 font names to alias the MIME names for the iso-8859-*
755 =item As a code reference, e.g.:
757 define_alias( sub { return /^iso8859-(\d+)$/i ? "iso-8859-$1" : undef } , '');
759 In this case C<$_> will be set to the name that is being looked up and
760 I<ENCODING> is passed to the sub as its first argument. The example
761 is another way to names as used in X11 font names to alias the MIME
762 names for the iso-8859-* family.
766 =head2 Defining Encodings
768 use Encode qw(define_alias);
769 define_encoding( $object, 'canonicalName' [,alias...]);
771 Causes I<canonicalName> to be associated with I<$object>. The object
772 should provide the interface described in L</"IMPLEMENTATION CLASSES">
773 below. If more than two arguments are provided then additional
774 arguments are taken as aliases for I<$object> as for C<define_alias>.
776 =head1 Encoding and IO
778 It is very common to want to do encoding transformations when
779 reading or writing files, network connections, pipes etc.
780 If Perl is configured to use the new 'perlio' IO system then
781 C<Encode> provides a "layer" (See L<perliol>) which can transform
782 data as it is read or written.
784 Here is how the blind poet would modernise the encoding:
787 open(my $iliad,'<:encoding(iso-8859-7)','iliad.greek');
788 open(my $utf8,'>:utf8','iliad.utf8');
794 In addition the new IO system can also be configured to read/write
795 UTF-8 encoded characters (as noted above this is efficient):
797 open(my $fh,'>:utf8','anything');
798 print $fh "Any \x{0021} string \N{SMILEY FACE}\n";
800 Either of the above forms of "layer" specifications can be made the default
801 for a lexical scope with the C<use open ...> pragma. See L<open>.
803 Once a handle is open is layers can be altered using C<binmode>.
805 Without any such configuration, or if Perl itself is built using
806 system's own IO, then write operations assume that file handle accepts
807 only I<bytes> and will C<die> if a character larger than 255 is
808 written to the handle. When reading, each octet from the handle
809 becomes a byte-in-a-character. Note that this default is the same
810 behaviour as bytes-only languages (including Perl before v5.6) would
811 have, and is sufficient to handle native 8-bit encodings
812 e.g. iso-8859-1, EBCDIC etc. and any legacy mechanisms for handling
813 other encodings and binary data.
815 In other cases it is the programs responsibility to transform
816 characters into bytes using the API above before doing writes, and to
817 transform the bytes read from a handle into characters before doing
818 "character operations" (e.g. C<lc>, C</\W+/>, ...).
820 You can also use PerlIO to convert larger amounts of data you don't
821 want to bring into memory. For example to convert between ISO 8859-1
822 (Latin 1) and UTF-8 (or UTF-EBCDIC in EBCDIC machines):
824 open(F, "<:encoding(iso-8859-1)", "data.txt") or die $!;
825 open(G, ">:utf8", "data.utf") or die $!;
826 while (<F>) { print G }
828 # Could also do "print G <F>" but that would pull
829 # the whole file into memory just to write it out again.
833 open(my $f, "<:encoding(cp1252)")
834 open(my $g, ">:encoding(iso-8859-2)")
835 open(my $h, ">:encoding(latin9)") # iso-8859-15
837 See L<PerlIO> for more information.
839 See also L<encoding> for how to change the default encoding of the
842 =head1 Encoding How to ...
848 =item * IO with mixed content (faking iso-2020-*)
850 =item * MIME's Content-Length:
852 =item * UTF-8 strings in binary data.
854 =item * Perl/Encode wrappers on non-Unicode XS modules.
858 =head1 Messing with Perl's Internals
860 The following API uses parts of Perl's internals in the current
861 implementation. As such they are efficient, but may change.
865 =item * is_utf8(STRING [, CHECK])
867 [INTERNAL] Test whether the UTF-8 flag is turned on in the STRING.
868 If CHECK is true, also checks the data in STRING for being well-formed
869 UTF-8. Returns true if successful, false otherwise.
871 =item * valid_utf8(STRING)
873 [INTERNAL] Test whether STRING is in a consistent state. Will return
874 true if string is held as bytes, or is well-formed UTF-8 and has the
875 UTF-8 flag on. Main reason for this routine is to allow Perl's
876 testsuite to check that operations have left strings in a consistent
883 [INTERNAL] Turn on the UTF-8 flag in STRING. The data in STRING is
884 B<not> checked for being well-formed UTF-8. Do not use unless you
885 B<know> that the STRING is well-formed UTF-8. Returns the previous
886 state of the UTF-8 flag (so please don't test the return value as
887 I<not> success or failure), or C<undef> if STRING is not a string.
893 [INTERNAL] Turn off the UTF-8 flag in STRING. Do not use frivolously.
894 Returns the previous state of the UTF-8 flag (so please don't test the
895 return value as I<not> success or failure), or C<undef> if STRING is
900 =head1 IMPLEMENTATION CLASSES
902 As mentioned above encodings are (in the current implementation at least)
903 defined by objects. The mapping of encoding name to object is via the
906 The values of the hash can currently be either strings or objects.
907 The string form may go away in the future. The string form occurs
908 when C<encodings()> has scanned C<@INC> for loadable encodings but has
909 not actually loaded the encoding in question. This is because the
910 current "loading" process is all Perl and a bit slow.
912 Once an encoding is loaded then value of the hash is object which
913 implements the encoding. The object should provide the following
920 Should return the string representing the canonical name of the encoding.
922 =item -E<gt>new_sequence
924 This is a placeholder for encodings with state. It should return an
925 object which implements this interface, all current implementations
926 return the original object.
928 =item -E<gt>encode($string,$check)
930 Should return the octet sequence representing I<$string>. If I<$check>
931 is true it should modify I<$string> in place to remove the converted
932 part (i.e. the whole string unless there is an error). If an error
933 occurs it should return the octet sequence for the fragment of string
934 that has been converted, and modify $string in-place to remove the
935 converted part leaving it starting with the problem fragment.
937 If check is is false then C<encode> should make a "best effort" to
938 convert the string - for example by using a replacement character.
940 =item -E<gt>decode($octets,$check)
942 Should return the string that I<$octets> represents. If I<$check> is
943 true it should modify I<$octets> in place to remove the converted part
944 (i.e. the whole sequence unless there is an error). If an error
945 occurs it should return the fragment of string that has been
946 converted, and modify $octets in-place to remove the converted part
947 leaving it starting with the problem fragment.
949 If check is is false then C<decode> should make a "best effort" to
950 convert the string - for example by using Unicode's "\x{FFFD}" as a
951 replacement character.
955 It should be noted that the check behaviour is different from the
956 outer public API. The logic is that the "unchecked" case is useful
957 when encoding is part of a stream which may be reporting errors
958 (e.g. STDERR). In such cases it is desirable to get everything
959 through somehow without causing additional errors which obscure the
960 original one. Also the encoding is best placed to know what the
961 correct replacement character is, so if that is the desired behaviour
962 then letting low level code do it is the most efficient.
964 In contrast if check is true, the scheme above allows the encoding to
965 do as much as it can and tell layer above how much that was. What is
966 lacking at present is a mechanism to report what went wrong. The most
967 likely interface will be an additional method call to the object, or
968 perhaps (to avoid forcing per-stream objects on otherwise stateless
969 encodings) and additional parameter.
971 It is also highly desirable that encoding classes inherit from
972 C<Encode::Encoding> as a base class. This allows that class to define
973 additional behaviour for all encoding objects. For example built in
974 Unicode, UCS-2 and UTF-8 classes use :
976 package Encode::MyEncoding;
977 use base qw(Encode::Encoding);
979 __PACKAGE__->Define(qw(myCanonical myAlias));
981 To create an object with bless {Name => ...},$class, and call
982 define_encoding. They inherit their C<name> method from
985 =head2 Compiled Encodings
987 F<Encode.xs> provides a class C<Encode::XS> which provides the
988 interface described above. It calls a generic octet-sequence to
989 octet-sequence "engine" that is driven by tables (defined in
990 F<encengine.c>). The same engine is used for both encode and
991 decode. C<Encode:XS>'s C<encode> forces Perl's characters to their
992 UTF-8 form and then treats them as just another multibyte
993 encoding. C<Encode:XS>'s C<decode> transforms the sequence and then
994 turns the UTF-8-ness flag as that is the form that the tables are
995 defined to produce. For details of the engine see the comments in
998 The tables are produced by the Perl script F<compile> (the name needs
999 to change so we can eventually install it somewhere). F<compile> can
1000 currently read two formats:
1006 This is a coined format used by Tcl. It is documented in
1007 Encode/EncodeFormat.pod.
1011 This is the semi-standard format used by IBM's ICU package.
1015 F<compile> can write the following forms:
1021 See above - the F<Encode/*.ucm> files provided with the distribution have
1022 been created from the original Tcl .enc files using this approach.
1026 Produces tables as C data structures - this is used to build in encodings
1027 into F<Encode.so>/F<Encode.dll>.
1031 In theory this allows encodings to be stand-alone loadable Perl
1032 extensions. The process has not yet been tested. The plan is to use
1033 this approach for large East Asian encodings.
1037 The set of encodings built-in to F<Encode.so>/F<Encode.dll> is
1038 determined by F<Makefile.PL>. The current set is as follows:
1042 =item ascii and iso-8859-*
1044 That is all the common 8-bit "western" encodings.
1046 =item IBM-1047 and two other variants of EBCDIC.
1048 These are the same variants that are supported by EBCDIC Perl as
1049 "native" encodings. They are included to prove "reversibility" of
1050 some constructs in EBCDIC Perl.
1052 =item symbol and dingbats as used by Tk on X11.
1054 (The reason Encode got started was to support Perl/Tk.)
1058 That set is rather ad hoc and has been driven by the needs of the
1059 tests rather than the needs of typical applications. It is likely
1064 L<perlunicode>, L<perlebcdic>, L<perlfunc/open>, L<PerlIO>, L<encoding>