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: Armenian encoding ARMSCII-8
177 # TODO: Hebrew encoding ISO-8859-8-1
178 # TODO: Thai encoding TCVN
179 # TODO: Korean encoding Johab
180 # TODO: Vietnamese encodings VPS
181 # TODO: Japanese encoding JIS (not the same as SJIS)
182 # TODO: Mac Asian+African encodings: Arabic Armenian Bengali Burmese
183 # ChineseSimp ChineseTrad Devanagari Ethiopic ExtArabic
184 # Farsi Georgian Gujarati Gurmukhi Hebrew Japanese
185 # Kannada Khmer Korean Laotian Malayalam Mongolian
186 # Oriya Sinhalese Symbol Tamil Telugu Tibetan Vietnamese
188 # Map white space and _ to '-'
189 define_alias( qr/^(\S+)[\s_]+(.*)$/i => '"$1-$2"' );
195 $encoding{$name} = $obj;
197 define_alias($lc => $obj) unless $lc eq $name;
201 define_alias($alias,$obj);
208 my ($class,$name) = @_;
210 if (ref($name) && $name->can('new_sequence'))
215 if (exists $encoding{$name})
217 return $encoding{$name};
219 if (exists $encoding{$lc})
221 return $encoding{$lc};
224 my $oc = $class->findAlias($name);
225 return $oc if defined $oc;
226 return $class->findAlias($lc) if $lc ne $name;
234 return __PACKAGE__->getEncoding($name);
239 my ($name,$string,$check) = @_;
240 my $enc = find_encoding($name);
241 croak("Unknown encoding '$name'") unless defined $enc;
242 my $octets = $enc->encode($string,$check);
243 return undef if ($check && length($string));
249 my ($name,$octets,$check) = @_;
250 my $enc = find_encoding($name);
251 croak("Unknown encoding '$name'") unless defined $enc;
252 my $string = $enc->decode($octets,$check);
253 $_[1] = $octets if $check;
259 my ($string,$from,$to,$check) = @_;
260 my $f = find_encoding($from);
261 croak("Unknown encoding '$from'") unless defined $f;
262 my $t = find_encoding($to);
263 croak("Unknown encoding '$to'") unless defined $t;
264 my $uni = $f->decode($string,$check);
265 return undef if ($check && length($string));
266 $string = $t->encode($uni,$check);
267 return undef if ($check && length($uni));
268 return length($_[0] = $string);
281 return undef unless utf8::decode($str);
285 require Encode::Encoding;
287 require Encode::Internal;
288 require Encode::Unicode;
289 require Encode::utf8;
290 require Encode::iso10646_1;
291 require Encode::ucs2_le;
299 Encode - character encodings
305 use Encode::TW; # for Taiwan-based Chinese encodings
306 use Encode::CN; # for China-based Chinese encodings
307 use Encode::JP; # for Japanese encodings
308 use Encode::KR; # for Korean encodings
312 The C<Encode> module provides the interfaces between Perl's strings
313 and the rest of the system. Perl strings are sequences of B<characters>.
315 The repertoire of characters that Perl can represent is at least that
316 defined by the Unicode Consortium. On most platforms the ordinal
317 values of the characters (as returned by C<ord(ch)>) is the "Unicode
318 codepoint" for the character (the exceptions are those platforms where
319 the legacy encoding is some variant of EBCDIC rather than a super-set
320 of ASCII - see L<perlebcdic>).
322 Traditionaly computer data has been moved around in 8-bit chunks
323 often called "bytes". These chunks are also known as "octets" in
324 networking standards. Perl is widely used to manipulate data of
325 many types - not only strings of characters representing human or
326 computer languages but also "binary" data being the machines representation
327 of numbers, pixels in an image - or just about anything.
329 When Perl is processing "binary data" the programmer wants Perl to process
330 "sequences of bytes". This is not a problem for Perl - as a byte has 256
331 possible values it easily fits in Perl's much larger "logical character".
333 Due to size concerns, before using B<CJK> (Chinese, Japanese & Korean)
334 encodings, you have to C<use> the corresponding
335 B<Encode::>(B<TW>|B<CN>|B<JP>|B<KR>) modules first.
343 I<character>: a character in the range 0..(2**32-1) (or more).
344 (What Perl's strings are made of.)
348 I<byte>: a character in the range 0..255
349 (A special case of a Perl character.)
353 I<octet>: 8 bits of data, with ordinal values 0..255
354 (Term for bytes passed to or from a non-Perl context, e.g. disk file.)
358 The marker [INTERNAL] marks Internal Implementation Details, in
359 general meant only for those who think they know what they are doing,
360 and such details may change in future releases.
364 =head2 Characteristics of an Encoding
366 An encoding has a "repertoire" of characters that it can represent,
367 and for each representable character there is at least one sequence of
368 octets that represents it.
370 =head2 Types of Encodings
372 Encodings can be divided into the following types:
376 =item * Fixed length 8-bit (or less) encodings.
378 Each character is a single octet so may have a repertoire of up to
379 256 characters. ASCII and iso-8859-* are typical examples.
381 =item * Fixed length 16-bit encodings
383 Each character is two octets so may have a repertoire of up to
384 65 536 characters. Unicode's UCS-2 is an example. Also used for
385 encodings for East Asian languages.
387 =item * Fixed length 32-bit encodings.
389 Not really very "encoded" encodings. The Unicode code points
390 are just represented as 4-octet integers. None the less because
391 different architectures use different representations of integers
392 (so called "endian") there at least two disctinct encodings.
394 =item * Multi-byte encodings
396 The number of octets needed to represent a character varies.
397 UTF-8 is a particularly complex but regular case of a multi-byte
398 encoding. Several East Asian countries use a multi-byte encoding
399 where 1-octet is used to cover western roman characters and Asian
400 characters get 2-octets.
401 (UTF-16 is strictly a multi-byte encoding taking either 2 or 4 octets
402 to represent a Unicode code point.)
404 =item * "Escape" encodings.
406 These encodings embed "escape sequences" into the octet sequence
407 which describe how the following octets are to be interpreted.
408 The iso-2022-* family is typical. Following the escape sequence
409 octets are encoded by an "embedded" encoding (which will be one
410 of the above types) until another escape sequence switches to
411 a different "embedded" encoding.
413 These schemes are very flexible and can handle mixed languages but are
414 very complex to process (and have state). No escape encodings are
415 implemented for Perl yet.
419 =head2 Specifying Encodings
421 Encodings can be specified to the API described below in two ways:
427 Encoding names are strings with characters taken from a restricted
428 repertoire. See L</"Encoding Names">.
430 =item 2. As an object
432 Encoding objects are returned by C<find_encoding($name)>.
436 =head2 Encoding Names
438 Encoding names are case insensitive. White space in names is ignored.
439 In addition an encoding may have aliases. Each encoding has one
440 "canonical" name. The "canonical" name is chosen from the names of
441 the encoding by picking the first in the following sequence:
445 =item * The MIME name as defined in IETF RFCs.
447 =item * The name in the IANA registry.
449 =item * The name used by the organization that defined it.
453 Because of all the alias issues, and because in the general case
454 encodings have state C<Encode> uses the encoding object internally
455 once an operation is in progress.
457 As of Perl 5.8.0, at least the following encodings are recognized
458 (the => marks aliases):
472 The ISO 8859 and KOI:
474 ISO 8859-1 ISO 8859-6 ISO 8859-11 KOI8-F
475 ISO 8859-2 ISO 8859-7 (12 doesn't exist) KOI8-R
476 ISO 8859-3 ISO 8859-8 ISO 8859-13 KOI8-U
477 ISO 8859-4 ISO 8859-9 ISO 8859-14
478 ISO 8859-5 ISO 8859-10 ISO 8859-15
481 Latin1 => 8859-1 Latin6 => 8859-10
482 Latin2 => 8859-2 Latin7 => 8859-13
483 Latin3 => 8859-3 Latin8 => 8859-14
484 Latin4 => 8859-4 Latin9 => 8859-15
485 Latin5 => 8859-9 Latin10 => 8859-16
494 The CJKV: Chinese, Japanese, Korean, Vietnamese:
496 ISO 2022 ISO 2022 JP-1 JIS 0201 GB 1988 Big5 EUC-CN
497 ISO 2022 CN ISO 2022 JP-2 JIS 0208 GB 2312 HZ EUC-JP
498 ISO 2022 JP ISO 2022 KR JIS 0210 GB 12345 CNS 11643 EUC-JP-0212
499 Shift-JIS GBK Big5-HKSCS EUC-KR
502 (Due to size concerns, additional Chinese encodings including C<GB 18030>,
503 C<EUC-TW> and C<BIG5PLUS> are distributed separately on CPAN, under the name
504 L<Encode::HanExtra>.)
508 CP37 CP852 CP861 CP866 CP949 CP1251 CP1256
509 CP424 CP855 CP862 CP869 CP950 CP1252 CP1257
510 CP737 CP856 CP863 CP874 CP1006 CP1253 CP1258
511 CP775 CP857 CP864 CP932 CP1047 CP1254
512 CP850 CP860 CP865 CP936 CP1250 CP1255
516 WinCyrillic => CP1251
518 WinTurkiskh => CP1254
522 WinVietnamese => CP1258
524 (All the CPI<NNN...> are available also as IBMI<NNN...>.)
528 MacCentralEuropean MacJapanese
530 MacCyrillic MacRomanian
533 MacIcelandic MacTurkish
544 =head1 PERL ENCODING API
546 =head2 Generic Encoding Interface
552 $bytes = encode(ENCODING, $string[, CHECK])
554 Encodes string from Perl's internal form into I<ENCODING> and returns
555 a sequence of octets. For CHECK see L</"Handling Malformed Data">.
557 For example to convert (internally UTF-8 encoded) Unicode data
560 $octets = encode("utf8", $unicode);
564 $string = decode(ENCODING, $bytes[, CHECK])
566 Decode sequence of octets assumed to be in I<ENCODING> into Perl's
567 internal form and returns the resulting string. For CHECK see
568 L</"Handling Malformed Data">.
570 For example to convert ISO 8859-1 data to UTF-8:
572 $utf8 = decode("latin1", $latin1);
576 from_to($string, FROM_ENCODING, TO_ENCODING[, CHECK])
578 Convert B<in-place> the data between two encodings. How did the data
579 in $string originally get to be in FROM_ENCODING? Either using
580 encode() or through PerlIO: See L</"Encoding and IO">. For CHECK
581 see L</"Handling Malformed Data">.
583 For example to convert ISO 8859-1 data to UTF-8:
585 from_to($data, "iso-8859-1", "utf-8");
587 and to convert it back:
589 from_to($data, "utf-8", "iso-8859-1");
591 Note that because the conversion happens in place, the data to be
592 converted cannot be a string constant, it must be a scalar variable.
596 =head2 Handling Malformed Data
598 If CHECK is not set, C<undef> is returned. If the data is supposed to
599 be UTF-8, an optional lexical warning (category utf8) is given. If
600 CHECK is true but not a code reference, dies.
602 It would desirable to have a way to indicate that transform should use
603 the encodings "replacement character" - no such mechanism is defined yet.
605 It is also planned to allow I<CHECK> to be a code reference.
607 This is not yet implemented as there are design issues with what its
608 arguments should be and how it returns its results.
614 Passed remaining fragment of string being processed.
615 Modifies it in place to remove bytes/characters it can understand
616 and returns a string used to represent them.
620 my $ch = substr($_[0],0,1,'');
621 return sprintf("\x{%02X}",ord($ch);
624 This scheme is close to how underlying C code for Encode works, but gives
625 the fixup routine very little context.
629 Passed original string, and an index into it of the problem area, and
630 output string so far. Appends what it will to output string and
631 returns new index into original string. For example:
634 # my ($s,$i,$d) = @_;
635 my $ch = substr($_[0],$_[1],1);
636 $_[2] .= sprintf("\x{%02X}",ord($ch);
640 This scheme gives maximal control to the fixup routine but is more
641 complicated to code, and may need internals of Encode to be tweaked to
642 keep original string intact.
648 Multiple return values rather than in-place modifications.
650 Index into the string could be pos($str) allowing s/\G...//.
656 The Unicode consortium defines the UTF-8 standard as a way of encoding
657 the entire Unicode repertiore as sequences of octets. This encoding is
658 expected to become very widespread. Perl can use this form internaly
659 to represent strings, so conversions to and from this form are
660 particularly efficient (as octets in memory do not have to change,
661 just the meta-data that tells Perl how to treat them).
667 $bytes = encode_utf8($string);
669 The characters that comprise string are encoded in Perl's superset of UTF-8
670 and the resulting octets returned as a sequence of bytes. All possible
671 characters have a UTF-8 representation so this function cannot fail.
675 $string = decode_utf8($bytes [,CHECK]);
677 The sequence of octets represented by $bytes is decoded from UTF-8
678 into a sequence of logical characters. Not all sequences of octets
679 form valid UTF-8 encodings, so it is possible for this call to fail.
680 For CHECK see L</"Handling Malformed Data">.
684 =head2 Other Encodings of Unicode
686 UTF-16 is similar to UCS-2, 16 bit or 2-byte chunks. UCS-2 can only
687 represent 0..0xFFFF, while UTF-16 has a I<surrogate pair> scheme which
688 allows it to cover the whole Unicode range.
690 Surrogates are code points set aside to encode the 0x01000..0x10FFFF
691 range of Unicode code points in pairs of 16-bit units. The I<high
692 surrogates> are the range 0xD800..0xDBFF, and the I<low surrogates>
693 are the range 0xDC00..0xDFFFF. The surrogate encoding is
695 $hi = ($uni - 0x10000) / 0x400 + 0xD800;
696 $lo = ($uni - 0x10000) % 0x400 + 0xDC00;
700 $uni = 0x10000 + ($hi - 0xD8000) * 0x400 + ($lo - 0xDC00);
702 Encode implements big-endian UCS-2 aliased to "iso-10646-1" as that
703 happens to be the name used by that representation when used with X11
706 UTF-32 or UCS-4 is 32-bit or 4-byte chunks. Perl's logical characters
707 can be considered as being in this form without encoding. An encoding
708 to transfer strings in this form (e.g. to write them to a file) would
711 pack('L*', unpack('U*', $string)); # native
713 pack('V*', unpack('U*', $string)); # little-endian
715 pack('N*', unpack('U*', $string)); # big-endian
717 depending on the endianness required.
719 No UTF-32 encodings are implemented yet.
721 Both UCS-2 and UCS-4 style encodings can have "byte order marks" by
722 representing the code point 0xFFFE as the very first thing in a file.
724 =head2 Listing available encodings
726 use Encode qw(encodings);
729 Returns a list of the canonical names of the available encodings.
731 =head2 Defining Aliases
733 use Encode qw(define_alias);
734 define_alias( newName => ENCODING);
736 Allows newName to be used as am alias for ENCODING. ENCODING may be
737 either the name of an encoding or and encoding object (as above).
739 Currently I<newName> can be specified in the following ways:
743 =item As a simple string.
745 =item As a qr// compiled regular expression, e.g.:
747 define_alias( qr/^iso8859-(\d+)$/i => '"iso-8859-$1"' );
749 In this case if I<ENCODING> is not a reference it is C<eval>-ed to
750 allow C<$1> etc. to be subsituted. The example is one way to names as
751 used in X11 font names to alias the MIME names for the iso-8859-*
754 =item As a code reference, e.g.:
756 define_alias( sub { return /^iso8859-(\d+)$/i ? "iso-8859-$1" : undef } , '');
758 In this case C<$_> will be set to the name that is being looked up and
759 I<ENCODING> is passed to the sub as its first argument. The example
760 is another way to names as used in X11 font names to alias the MIME
761 names for the iso-8859-* family.
765 =head2 Defining Encodings
767 use Encode qw(define_alias);
768 define_encoding( $object, 'canonicalName' [,alias...]);
770 Causes I<canonicalName> to be associated with I<$object>. The object
771 should provide the interface described in L</"IMPLEMENTATION CLASSES">
772 below. If more than two arguments are provided then additional
773 arguments are taken as aliases for I<$object> as for C<define_alias>.
775 =head1 Encoding and IO
777 It is very common to want to do encoding transformations when
778 reading or writing files, network connections, pipes etc.
779 If Perl is configured to use the new 'perlio' IO system then
780 C<Encode> provides a "layer" (See L<perliol>) which can transform
781 data as it is read or written.
783 Here is how the blind poet would modernise the encoding:
786 open(my $iliad,'<:encoding(iso-8859-7)','iliad.greek');
787 open(my $utf8,'>:utf8','iliad.utf8');
793 In addition the new IO system can also be configured to read/write
794 UTF-8 encoded characters (as noted above this is efficient):
796 open(my $fh,'>:utf8','anything');
797 print $fh "Any \x{0021} string \N{SMILEY FACE}\n";
799 Either of the above forms of "layer" specifications can be made the default
800 for a lexical scope with the C<use open ...> pragma. See L<open>.
802 Once a handle is open is layers can be altered using C<binmode>.
804 Without any such configuration, or if Perl itself is built using
805 system's own IO, then write operations assume that file handle accepts
806 only I<bytes> and will C<die> if a character larger than 255 is
807 written to the handle. When reading, each octet from the handle
808 becomes a byte-in-a-character. Note that this default is the same
809 behaviour as bytes-only languages (including Perl before v5.6) would
810 have, and is sufficient to handle native 8-bit encodings
811 e.g. iso-8859-1, EBCDIC etc. and any legacy mechanisms for handling
812 other encodings and binary data.
814 In other cases it is the programs responsibility to transform
815 characters into bytes using the API above before doing writes, and to
816 transform the bytes read from a handle into characters before doing
817 "character operations" (e.g. C<lc>, C</\W+/>, ...).
819 You can also use PerlIO to convert larger amounts of data you don't
820 want to bring into memory. For example to convert between ISO 8859-1
821 (Latin 1) and UTF-8 (or UTF-EBCDIC in EBCDIC machines):
823 open(F, "<:encoding(iso-8859-1)", "data.txt") or die $!;
824 open(G, ">:utf8", "data.utf") or die $!;
825 while (<F>) { print G }
827 # Could also do "print G <F>" but that would pull
828 # the whole file into memory just to write it out again.
832 open(my $f, "<:encoding(cp1252)")
833 open(my $g, ">:encoding(iso-8859-2)")
834 open(my $h, ">:encoding(latin9)") # iso-8859-15
836 See L<PerlIO> for more information.
838 See also L<encoding> for how to change the default encoding of the
841 =head1 Encoding How to ...
847 =item * IO with mixed content (faking iso-2020-*)
849 =item * MIME's Content-Length:
851 =item * UTF-8 strings in binary data.
853 =item * Perl/Encode wrappers on non-Unicode XS modules.
857 =head1 Messing with Perl's Internals
859 The following API uses parts of Perl's internals in the current
860 implementation. As such they are efficient, but may change.
864 =item * is_utf8(STRING [, CHECK])
866 [INTERNAL] Test whether the UTF-8 flag is turned on in the STRING.
867 If CHECK is true, also checks the data in STRING for being well-formed
868 UTF-8. Returns true if successful, false otherwise.
870 =item * valid_utf8(STRING)
872 [INTERNAL] Test whether STRING is in a consistent state. Will return
873 true if string is held as bytes, or is well-formed UTF-8 and has the
874 UTF-8 flag on. Main reason for this routine is to allow Perl's
875 testsuite to check that operations have left strings in a consistent
882 [INTERNAL] Turn on the UTF-8 flag in STRING. The data in STRING is
883 B<not> checked for being well-formed UTF-8. Do not use unless you
884 B<know> that the STRING is well-formed UTF-8. Returns the previous
885 state of the UTF-8 flag (so please don't test the return value as
886 I<not> success or failure), or C<undef> if STRING is not a string.
892 [INTERNAL] Turn off the UTF-8 flag in STRING. Do not use frivolously.
893 Returns the previous state of the UTF-8 flag (so please don't test the
894 return value as I<not> success or failure), or C<undef> if STRING is
899 =head1 IMPLEMENTATION CLASSES
901 As mentioned above encodings are (in the current implementation at least)
902 defined by objects. The mapping of encoding name to object is via the
905 The values of the hash can currently be either strings or objects.
906 The string form may go away in the future. The string form occurs
907 when C<encodings()> has scanned C<@INC> for loadable encodings but has
908 not actually loaded the encoding in question. This is because the
909 current "loading" process is all Perl and a bit slow.
911 Once an encoding is loaded then value of the hash is object which
912 implements the encoding. The object should provide the following
919 Should return the string representing the canonical name of the encoding.
921 =item -E<gt>new_sequence
923 This is a placeholder for encodings with state. It should return an
924 object which implements this interface, all current implementations
925 return the original object.
927 =item -E<gt>encode($string,$check)
929 Should return the octet sequence representing I<$string>. If I<$check>
930 is true it should modify I<$string> in place to remove the converted
931 part (i.e. the whole string unless there is an error). If an error
932 occurs it should return the octet sequence for the fragment of string
933 that has been converted, and modify $string in-place to remove the
934 converted part leaving it starting with the problem fragment.
936 If check is is false then C<encode> should make a "best effort" to
937 convert the string - for example by using a replacement character.
939 =item -E<gt>decode($octets,$check)
941 Should return the string that I<$octets> represents. If I<$check> is
942 true it should modify I<$octets> in place to remove the converted part
943 (i.e. the whole sequence unless there is an error). If an error
944 occurs it should return the fragment of string that has been
945 converted, and modify $octets in-place to remove the converted part
946 leaving it starting with the problem fragment.
948 If check is is false then C<decode> should make a "best effort" to
949 convert the string - for example by using Unicode's "\x{FFFD}" as a
950 replacement character.
954 It should be noted that the check behaviour is different from the
955 outer public API. The logic is that the "unchecked" case is useful
956 when encoding is part of a stream which may be reporting errors
957 (e.g. STDERR). In such cases it is desirable to get everything
958 through somehow without causing additional errors which obscure the
959 original one. Also the encoding is best placed to know what the
960 correct replacement character is, so if that is the desired behaviour
961 then letting low level code do it is the most efficient.
963 In contrast if check is true, the scheme above allows the encoding to
964 do as much as it can and tell layer above how much that was. What is
965 lacking at present is a mechanism to report what went wrong. The most
966 likely interface will be an additional method call to the object, or
967 perhaps (to avoid forcing per-stream objects on otherwise stateless
968 encodings) and additional parameter.
970 It is also highly desirable that encoding classes inherit from
971 C<Encode::Encoding> as a base class. This allows that class to define
972 additional behaviour for all encoding objects. For example built in
973 Unicode, UCS-2 and UTF-8 classes use :
975 package Encode::MyEncoding;
976 use base qw(Encode::Encoding);
978 __PACKAGE__->Define(qw(myCanonical myAlias));
980 To create an object with bless {Name => ...},$class, and call
981 define_encoding. They inherit their C<name> method from
984 =head2 Compiled Encodings
986 F<Encode.xs> provides a class C<Encode::XS> which provides the
987 interface described above. It calls a generic octet-sequence to
988 octet-sequence "engine" that is driven by tables (defined in
989 F<encengine.c>). The same engine is used for both encode and
990 decode. C<Encode:XS>'s C<encode> forces Perl's characters to their
991 UTF-8 form and then treats them as just another multibyte
992 encoding. C<Encode:XS>'s C<decode> transforms the sequence and then
993 turns the UTF-8-ness flag as that is the form that the tables are
994 defined to produce. For details of the engine see the comments in
997 The tables are produced by the Perl script F<compile> (the name needs
998 to change so we can eventually install it somewhere). F<compile> can
999 currently read two formats:
1005 This is a coined format used by Tcl. It is documented in
1006 Encode/EncodeFormat.pod.
1010 This is the semi-standard format used by IBM's ICU package.
1014 F<compile> can write the following forms:
1020 See above - the F<Encode/*.ucm> files provided with the distribution have
1021 been created from the original Tcl .enc files using this approach.
1025 Produces tables as C data structures - this is used to build in encodings
1026 into F<Encode.so>/F<Encode.dll>.
1030 In theory this allows encodings to be stand-alone loadable Perl
1031 extensions. The process has not yet been tested. The plan is to use
1032 this approach for large East Asian encodings.
1036 The set of encodings built-in to F<Encode.so>/F<Encode.dll> is
1037 determined by F<Makefile.PL>. The current set is as follows:
1041 =item ascii and iso-8859-*
1043 That is all the common 8-bit "western" encodings.
1045 =item IBM-1047 and two other variants of EBCDIC.
1047 These are the same variants that are supported by EBCDIC Perl as
1048 "native" encodings. They are included to prove "reversibility" of
1049 some constructs in EBCDIC Perl.
1051 =item symbol and dingbats as used by Tk on X11.
1053 (The reason Encode got started was to support Perl/Tk.)
1057 That set is rather ad hoc and has been driven by the needs of the
1058 tests rather than the needs of typical applications. It is likely
1063 L<perlunicode>, L<perlebcdic>, L<perlfunc/open>, L<PerlIO>, L<encoding>