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 );
60 our %external_tables =
62 'euc-cn' => 'Encode/CN.pm',
63 gb2312 => 'Encode/CN.pm',
64 gb12345 => 'Encode/CN.pm',
65 gbk => 'Encode/CN.pm',
66 cp936 => 'Encode/CN.pm',
67 'iso-ir-165' => 'Encode/CN.pm',
68 'euc-jp' => 'Encode/JP.pm',
69 shiftjis => 'Encode/JP.pm',
70 macjapan => 'Encode/JP.pm',
71 cp932 => 'Encode/JP.pm',
72 'euc-kr' => 'Encode/KR.pm',
73 ksc5601 => 'Encode/KR.pm',
74 cp949 => 'Encode/KR.pm',
75 big5 => 'Encode/TW.pm',
76 'big5-hkscs' => 'Encode/TW.pm',
77 cp950 => 'Encode/TW.pm',
78 gb18030 => 'Encode/HanExtra.pm',
79 big5plus => 'Encode/HanExtra.pm',
80 'euc-tw' => 'Encode/HanExtra.pm',
88 sort { $a->[1] cmp $b->[1] }
90 grep { $_ ne 'Internal' }
98 # print "# findAlias $_\n";
99 unless (exists $alias{$_})
101 for (my $i=0; $i < @alias; $i += 2)
103 my $alias = $alias[$i];
104 my $val = $alias[$i+1];
106 if (ref($alias) eq 'Regexp' && $_ =~ $alias)
110 elsif (ref($alias) eq 'CODE')
112 $new = &{$alias}($val)
114 elsif (lc($_) eq lc($alias))
120 next if $new eq $_; # avoid (direct) recursion on bugs
121 my $enc = (ref($new)) ? $new : find_encoding($new);
137 my ($alias,$name) = splice(@_,0,2);
138 push(@alias, $alias => $name);
142 # Allow variants of iso-8859-1 etc.
143 define_alias( qr/^iso[-_]?(\d+)[-_](\d+)$/i => '"iso-$1-$2"' );
145 # At least HP-UX has these.
146 define_alias( qr/^iso8859(\d+)$/i => '"iso-8859-$1"' );
149 define_alias( qr/^(?:hp-)?(arabic|greek|hebrew|kana|roman|thai|turkish)8$/i => '"${1}8"' );
151 # The Official name of ASCII.
152 define_alias( qr/^ANSI[-_]?X3\.4[-_]?1968$/i => '"ascii"' );
154 # This is a font issue, not an encoding issue.
155 # (The currency symbol of the Latin 1 upper half
156 # has been redefined as the euro symbol.)
157 define_alias( qr/^(.+)\@euro$/i => '"$1"' );
159 # Allow latin-1 style names as well
160 define_alias( qr/^(?:iso[-_]?)?latin[-_]?(\d+)$/i => '"iso-8859-$latin2iso_num[$1]"' );
162 # Allow winlatin1 style names as well
163 define_alias( qr/^win(latin[12]|cyrillic|baltic|greek|turkish|hebrew|arabic|baltic|vietnamese)$/i => '"cp$winlatin2cp{\u$1}"' );
165 # Common names for non-latin prefered MIME names
166 define_alias( 'ascii' => 'US-ascii',
167 'cyrillic' => 'iso-8859-5',
168 'arabic' => 'iso-8859-6',
169 'greek' => 'iso-8859-7',
170 'hebrew' => 'iso-8859-8',
171 'thai' => 'iso-8859-11',
172 'tis620' => 'iso-8859-11',
175 # At least AIX has IBM-NNN (surprisingly...) instead of cpNNN.
176 # And Microsoft has their own naming (again, surprisingly).
177 define_alias( qr/^(?:ibm|ms)[-_]?(\d\d\d\d?)$/i => '"cp$1"');
179 # Sometimes seen with a leading zero.
180 define_alias( qr/^cp037$/i => '"cp37"');
183 define_alias( qr/^macRomanian$/i => '"macRumanian"');
185 # Standardize on the dashed versions.
186 define_alias( qr/^utf8$/i => 'utf-8' );
187 define_alias( qr/^koi8r$/i => 'koi8-r' );
188 define_alias( qr/^koi8u$/i => 'koi8-u' );
190 # Seen in some Linuxes.
191 define_alias( qr/^ujis$/i => 'euc-jp' );
193 # CP936 doesn't have vendor-addon for GBK, so they're identical.
194 define_alias( qr/^gbk$/i => '"cp936"');
196 # TODO: HP-UX '8' encodings arabic8 greek8 hebrew8 kana8 thai8 turkish8
197 # TODO: HP-UX '15' encodings japanese15 korean15 roi15
198 # TODO: Cyrillic encoding ISO-IR-111 (useful?)
199 # TODO: Armenian encoding ARMSCII-8
200 # TODO: Hebrew encoding ISO-8859-8-1
201 # TODO: Thai encoding TCVN
202 # TODO: Korean encoding Johab
203 # TODO: Vietnamese encodings VPS
204 # TODO: Japanese encoding JIS (not the same as SJIS)
205 # TODO: Mac Asian+African encodings: Arabic Armenian Bengali Burmese
206 # ChineseSimp ChineseTrad Devanagari Ethiopic ExtArabic
207 # Farsi Georgian Gujarati Gurmukhi Hebrew Japanese
208 # Kannada Khmer Korean Laotian Malayalam Mongolian
209 # Oriya Sinhalese Symbol Tamil Telugu Tibetan Vietnamese
211 # Map white space and _ to '-'
212 define_alias( qr/^(\S+)[\s_]+(.*)$/i => '"$1-$2"' );
218 $encoding{$name} = $obj;
220 define_alias($lc => $obj) unless $lc eq $name;
224 define_alias($alias,$obj);
231 my ($class,$name) = @_;
233 if (ref($name) && $name->can('new_sequence'))
238 if (exists $encoding{$name})
240 return $encoding{$name};
242 if (exists $encoding{$lc})
244 return $encoding{$lc};
246 if (exists $external_tables{$lc})
248 require $external_tables{$lc};
249 return $encoding{$name} if exists $encoding{$name};
252 my $oc = $class->findAlias($name);
253 return $oc if defined $oc;
254 return $class->findAlias($lc) if $lc ne $name;
262 return __PACKAGE__->getEncoding($name);
267 my ($name,$string,$check) = @_;
268 my $enc = find_encoding($name);
269 croak("Unknown encoding '$name'") unless defined $enc;
270 my $octets = $enc->encode($string,$check);
271 return undef if ($check && length($string));
277 my ($name,$octets,$check) = @_;
278 my $enc = find_encoding($name);
279 croak("Unknown encoding '$name'") unless defined $enc;
280 my $string = $enc->decode($octets,$check);
281 $_[1] = $octets if $check;
287 my ($string,$from,$to,$check) = @_;
288 my $f = find_encoding($from);
289 croak("Unknown encoding '$from'") unless defined $f;
290 my $t = find_encoding($to);
291 croak("Unknown encoding '$to'") unless defined $t;
292 my $uni = $f->decode($string,$check);
293 return undef if ($check && length($string));
294 $string = $t->encode($uni,$check);
295 return undef if ($check && length($uni));
296 return length($_[0] = $string);
309 return undef unless utf8::decode($str);
313 require Encode::Encoding;
315 require Encode::Internal;
316 require Encode::Unicode;
317 require Encode::utf8;
318 require Encode::iso10646_1;
319 require Encode::ucs2_le;
327 Encode - character encodings
335 The C<Encode> module provides the interfaces between Perl's strings
336 and the rest of the system. Perl strings are sequences of B<characters>.
338 The repertoire of characters that Perl can represent is at least that
339 defined by the Unicode Consortium. On most platforms the ordinal
340 values of the characters (as returned by C<ord(ch)>) is the "Unicode
341 codepoint" for the character (the exceptions are those platforms where
342 the legacy encoding is some variant of EBCDIC rather than a super-set
343 of ASCII - see L<perlebcdic>).
345 Traditionaly computer data has been moved around in 8-bit chunks
346 often called "bytes". These chunks are also known as "octets" in
347 networking standards. Perl is widely used to manipulate data of
348 many types - not only strings of characters representing human or
349 computer languages but also "binary" data being the machines representation
350 of numbers, pixels in an image - or just about anything.
352 When Perl is processing "binary data" the programmer wants Perl to process
353 "sequences of bytes". This is not a problem for Perl - as a byte has 256
354 possible values it easily fits in Perl's much larger "logical character".
356 Due to size concerns, each of B<CJK> (Chinese, Japanese & Korean) modules
357 are not loaded in memory until the first time they're used. Although you
358 don't have to C<use> the corresponding B<Encode::>(B<TW>|B<CN>|B<JP>|B<KR>)
359 modules first, be aware that those encodings will not be in C<%encodings>
360 until their module is loaded (either implicitly through using encodings
361 contained in the same module, or via an explicit C<use>).
369 I<character>: a character in the range 0..(2**32-1) (or more).
370 (What Perl's strings are made of.)
374 I<byte>: a character in the range 0..255
375 (A special case of a Perl character.)
379 I<octet>: 8 bits of data, with ordinal values 0..255
380 (Term for bytes passed to or from a non-Perl context, e.g. disk file.)
384 The marker [INTERNAL] marks Internal Implementation Details, in
385 general meant only for those who think they know what they are doing,
386 and such details may change in future releases.
390 =head2 Characteristics of an Encoding
392 An encoding has a "repertoire" of characters that it can represent,
393 and for each representable character there is at least one sequence of
394 octets that represents it.
396 =head2 Types of Encodings
398 Encodings can be divided into the following types:
402 =item * Fixed length 8-bit (or less) encodings.
404 Each character is a single octet so may have a repertoire of up to
405 256 characters. ASCII and iso-8859-* are typical examples.
407 =item * Fixed length 16-bit encodings
409 Each character is two octets so may have a repertoire of up to
410 65 536 characters. Unicode's UCS-2 is an example. Also used for
411 encodings for East Asian languages.
413 =item * Fixed length 32-bit encodings.
415 Not really very "encoded" encodings. The Unicode code points
416 are just represented as 4-octet integers. None the less because
417 different architectures use different representations of integers
418 (so called "endian") there at least two disctinct encodings.
420 =item * Multi-byte encodings
422 The number of octets needed to represent a character varies.
423 UTF-8 is a particularly complex but regular case of a multi-byte
424 encoding. Several East Asian countries use a multi-byte encoding
425 where 1-octet is used to cover western roman characters and Asian
426 characters get 2-octets.
427 (UTF-16 is strictly a multi-byte encoding taking either 2 or 4 octets
428 to represent a Unicode code point.)
430 =item * "Escape" encodings.
432 These encodings embed "escape sequences" into the octet sequence
433 which describe how the following octets are to be interpreted.
434 The iso-2022-* family is typical. Following the escape sequence
435 octets are encoded by an "embedded" encoding (which will be one
436 of the above types) until another escape sequence switches to
437 a different "embedded" encoding.
439 These schemes are very flexible and can handle mixed languages but are
440 very complex to process (and have state). No escape encodings are
441 implemented for Perl yet.
445 =head2 Specifying Encodings
447 Encodings can be specified to the API described below in two ways:
453 Encoding names are strings with characters taken from a restricted
454 repertoire. See L</"Encoding Names">.
456 =item 2. As an object
458 Encoding objects are returned by C<find_encoding($name)>.
462 =head2 Encoding Names
464 Encoding names are case insensitive. White space in names is ignored.
465 In addition an encoding may have aliases. Each encoding has one
466 "canonical" name. The "canonical" name is chosen from the names of
467 the encoding by picking the first in the following sequence:
471 =item * The MIME name as defined in IETF RFCs.
473 =item * The name in the IANA registry.
475 =item * The name used by the organization that defined it.
479 Because of all the alias issues, and because in the general case
480 encodings have state C<Encode> uses the encoding object internally
481 once an operation is in progress.
483 As of Perl 5.8.0, at least the following encodings are recognized
484 (the => marks aliases):
498 The ISO 8859 and KOI:
500 ISO 8859-1 ISO 8859-6 ISO 8859-11 KOI8-F
501 ISO 8859-2 ISO 8859-7 (12 doesn't exist) KOI8-R
502 ISO 8859-3 ISO 8859-8 ISO 8859-13 KOI8-U
503 ISO 8859-4 ISO 8859-9 ISO 8859-14
504 ISO 8859-5 ISO 8859-10 ISO 8859-15
507 Latin1 => 8859-1 Latin6 => 8859-10
508 Latin2 => 8859-2 Latin7 => 8859-13
509 Latin3 => 8859-3 Latin8 => 8859-14
510 Latin4 => 8859-4 Latin9 => 8859-15
511 Latin5 => 8859-9 Latin10 => 8859-16
520 The CJKV: Chinese, Japanese, Korean, Vietnamese:
522 ISO 2022 ISO 2022 JP-1 JIS 0201 GB 1988 Big5 EUC-CN
523 ISO 2022 CN ISO 2022 JP-2 JIS 0208 GB 2312 HZ EUC-JP
524 ISO 2022 JP ISO 2022 KR JIS 0210 GB 12345 CNS 11643 EUC-JP-0212
525 Shift-JIS GBK Big5-HKSCS EUC-KR
528 (Due to size concerns, additional Chinese encodings including C<GB 18030>,
529 C<EUC-TW> and C<BIG5PLUS> are distributed separately on CPAN, under the name
530 L<Encode::HanExtra>.)
534 CP37 CP852 CP861 CP866 CP949 CP1251 CP1256
535 CP424 CP855 CP862 CP869 CP950 CP1252 CP1257
536 CP737 CP856 CP863 CP874 CP1006 CP1253 CP1258
537 CP775 CP857 CP864 CP932 CP1047 CP1254
538 CP850 CP860 CP865 CP936 CP1250 CP1255
542 WinCyrillic => CP1251
544 WinTurkiskh => CP1254
548 WinVietnamese => CP1258
550 (All the CPI<NNN...> are available also as IBMI<NNN...>.)
554 MacCentralEuropean MacJapanese
556 MacCyrillic MacRomanian
559 MacIcelandic MacTurkish
570 =head1 PERL ENCODING API
572 =head2 Generic Encoding Interface
578 $bytes = encode(ENCODING, $string[, CHECK])
580 Encodes string from Perl's internal form into I<ENCODING> and returns
581 a sequence of octets. For CHECK see L</"Handling Malformed Data">.
583 For example to convert (internally UTF-8 encoded) Unicode data
586 $octets = encode("utf8", $unicode);
590 $string = decode(ENCODING, $bytes[, CHECK])
592 Decode sequence of octets assumed to be in I<ENCODING> into Perl's
593 internal form and returns the resulting string. For CHECK see
594 L</"Handling Malformed Data">.
596 For example to convert ISO 8859-1 data to UTF-8:
598 $utf8 = decode("latin1", $latin1);
602 from_to($string, FROM_ENCODING, TO_ENCODING[, CHECK])
604 Convert B<in-place> the data between two encodings. How did the data
605 in $string originally get to be in FROM_ENCODING? Either using
606 encode() or through PerlIO: See L</"Encoding and IO">. For CHECK
607 see L</"Handling Malformed Data">.
609 For example to convert ISO 8859-1 data to UTF-8:
611 from_to($data, "iso-8859-1", "utf-8");
613 and to convert it back:
615 from_to($data, "utf-8", "iso-8859-1");
617 Note that because the conversion happens in place, the data to be
618 converted cannot be a string constant, it must be a scalar variable.
622 =head2 Handling Malformed Data
624 If CHECK is not set, C<undef> is returned. If the data is supposed to
625 be UTF-8, an optional lexical warning (category utf8) is given. If
626 CHECK is true but not a code reference, dies.
628 It would desirable to have a way to indicate that transform should use
629 the encodings "replacement character" - no such mechanism is defined yet.
631 It is also planned to allow I<CHECK> to be a code reference.
633 This is not yet implemented as there are design issues with what its
634 arguments should be and how it returns its results.
640 Passed remaining fragment of string being processed.
641 Modifies it in place to remove bytes/characters it can understand
642 and returns a string used to represent them.
646 my $ch = substr($_[0],0,1,'');
647 return sprintf("\x{%02X}",ord($ch);
650 This scheme is close to how underlying C code for Encode works, but gives
651 the fixup routine very little context.
655 Passed original string, and an index into it of the problem area, and
656 output string so far. Appends what it will to output string and
657 returns new index into original string. For example:
660 # my ($s,$i,$d) = @_;
661 my $ch = substr($_[0],$_[1],1);
662 $_[2] .= sprintf("\x{%02X}",ord($ch);
666 This scheme gives maximal control to the fixup routine but is more
667 complicated to code, and may need internals of Encode to be tweaked to
668 keep original string intact.
674 Multiple return values rather than in-place modifications.
676 Index into the string could be pos($str) allowing s/\G...//.
682 The Unicode consortium defines the UTF-8 standard as a way of encoding
683 the entire Unicode repertiore as sequences of octets. This encoding is
684 expected to become very widespread. Perl can use this form internaly
685 to represent strings, so conversions to and from this form are
686 particularly efficient (as octets in memory do not have to change,
687 just the meta-data that tells Perl how to treat them).
693 $bytes = encode_utf8($string);
695 The characters that comprise string are encoded in Perl's superset of UTF-8
696 and the resulting octets returned as a sequence of bytes. All possible
697 characters have a UTF-8 representation so this function cannot fail.
701 $string = decode_utf8($bytes [,CHECK]);
703 The sequence of octets represented by $bytes is decoded from UTF-8
704 into a sequence of logical characters. Not all sequences of octets
705 form valid UTF-8 encodings, so it is possible for this call to fail.
706 For CHECK see L</"Handling Malformed Data">.
710 =head2 Other Encodings of Unicode
712 UTF-16 is similar to UCS-2, 16 bit or 2-byte chunks. UCS-2 can only
713 represent 0..0xFFFF, while UTF-16 has a I<surrogate pair> scheme which
714 allows it to cover the whole Unicode range.
716 Surrogates are code points set aside to encode the 0x01000..0x10FFFF
717 range of Unicode code points in pairs of 16-bit units. The I<high
718 surrogates> are the range 0xD800..0xDBFF, and the I<low surrogates>
719 are the range 0xDC00..0xDFFFF. The surrogate encoding is
721 $hi = ($uni - 0x10000) / 0x400 + 0xD800;
722 $lo = ($uni - 0x10000) % 0x400 + 0xDC00;
726 $uni = 0x10000 + ($hi - 0xD8000) * 0x400 + ($lo - 0xDC00);
728 Encode implements big-endian UCS-2 aliased to "iso-10646-1" as that
729 happens to be the name used by that representation when used with X11
732 UTF-32 or UCS-4 is 32-bit or 4-byte chunks. Perl's logical characters
733 can be considered as being in this form without encoding. An encoding
734 to transfer strings in this form (e.g. to write them to a file) would
737 pack('L*', unpack('U*', $string)); # native
739 pack('V*', unpack('U*', $string)); # little-endian
741 pack('N*', unpack('U*', $string)); # big-endian
743 depending on the endianness required.
745 No UTF-32 encodings are implemented yet.
747 Both UCS-2 and UCS-4 style encodings can have "byte order marks" by
748 representing the code point 0xFFFE as the very first thing in a file.
750 =head2 Listing available encodings
752 use Encode qw(encodings);
755 Returns a list of the canonical names of the available encodings.
757 =head2 Defining Aliases
759 use Encode qw(define_alias);
760 define_alias( newName => ENCODING);
762 Allows newName to be used as am alias for ENCODING. ENCODING may be
763 either the name of an encoding or and encoding object (as above).
765 Currently I<newName> can be specified in the following ways:
769 =item As a simple string.
771 =item As a qr// compiled regular expression, e.g.:
773 define_alias( qr/^iso8859-(\d+)$/i => '"iso-8859-$1"' );
775 In this case if I<ENCODING> is not a reference it is C<eval>-ed to
776 allow C<$1> etc. to be subsituted. The example is one way to names as
777 used in X11 font names to alias the MIME names for the iso-8859-*
780 =item As a code reference, e.g.:
782 define_alias( sub { return /^iso8859-(\d+)$/i ? "iso-8859-$1" : undef } , '');
784 In this case C<$_> will be set to the name that is being looked up and
785 I<ENCODING> is passed to the sub as its first argument. The example
786 is another way to names as used in X11 font names to alias the MIME
787 names for the iso-8859-* family.
791 =head2 Defining Encodings
793 use Encode qw(define_alias);
794 define_encoding( $object, 'canonicalName' [,alias...]);
796 Causes I<canonicalName> to be associated with I<$object>. The object
797 should provide the interface described in L</"IMPLEMENTATION CLASSES">
798 below. If more than two arguments are provided then additional
799 arguments are taken as aliases for I<$object> as for C<define_alias>.
801 =head1 Encoding and IO
803 It is very common to want to do encoding transformations when
804 reading or writing files, network connections, pipes etc.
805 If Perl is configured to use the new 'perlio' IO system then
806 C<Encode> provides a "layer" (See L<perliol>) which can transform
807 data as it is read or written.
809 Here is how the blind poet would modernise the encoding:
812 open(my $iliad,'<:encoding(iso-8859-7)','iliad.greek');
813 open(my $utf8,'>:utf8','iliad.utf8');
819 In addition the new IO system can also be configured to read/write
820 UTF-8 encoded characters (as noted above this is efficient):
822 open(my $fh,'>:utf8','anything');
823 print $fh "Any \x{0021} string \N{SMILEY FACE}\n";
825 Either of the above forms of "layer" specifications can be made the default
826 for a lexical scope with the C<use open ...> pragma. See L<open>.
828 Once a handle is open is layers can be altered using C<binmode>.
830 Without any such configuration, or if Perl itself is built using
831 system's own IO, then write operations assume that file handle accepts
832 only I<bytes> and will C<die> if a character larger than 255 is
833 written to the handle. When reading, each octet from the handle
834 becomes a byte-in-a-character. Note that this default is the same
835 behaviour as bytes-only languages (including Perl before v5.6) would
836 have, and is sufficient to handle native 8-bit encodings
837 e.g. iso-8859-1, EBCDIC etc. and any legacy mechanisms for handling
838 other encodings and binary data.
840 In other cases it is the programs responsibility to transform
841 characters into bytes using the API above before doing writes, and to
842 transform the bytes read from a handle into characters before doing
843 "character operations" (e.g. C<lc>, C</\W+/>, ...).
845 You can also use PerlIO to convert larger amounts of data you don't
846 want to bring into memory. For example to convert between ISO 8859-1
847 (Latin 1) and UTF-8 (or UTF-EBCDIC in EBCDIC machines):
849 open(F, "<:encoding(iso-8859-1)", "data.txt") or die $!;
850 open(G, ">:utf8", "data.utf") or die $!;
851 while (<F>) { print G }
853 # Could also do "print G <F>" but that would pull
854 # the whole file into memory just to write it out again.
858 open(my $f, "<:encoding(cp1252)")
859 open(my $g, ">:encoding(iso-8859-2)")
860 open(my $h, ">:encoding(latin9)") # iso-8859-15
862 See L<PerlIO> for more information.
864 See also L<encoding> for how to change the default encoding of the
867 =head1 Encoding How to ...
873 =item * IO with mixed content (faking iso-2020-*)
875 =item * MIME's Content-Length:
877 =item * UTF-8 strings in binary data.
879 =item * Perl/Encode wrappers on non-Unicode XS modules.
883 =head1 Messing with Perl's Internals
885 The following API uses parts of Perl's internals in the current
886 implementation. As such they are efficient, but may change.
890 =item * is_utf8(STRING [, CHECK])
892 [INTERNAL] Test whether the UTF-8 flag is turned on in the STRING.
893 If CHECK is true, also checks the data in STRING for being well-formed
894 UTF-8. Returns true if successful, false otherwise.
896 =item * valid_utf8(STRING)
898 [INTERNAL] Test whether STRING is in a consistent state. Will return
899 true if string is held as bytes, or is well-formed UTF-8 and has the
900 UTF-8 flag on. Main reason for this routine is to allow Perl's
901 testsuite to check that operations have left strings in a consistent
908 [INTERNAL] Turn on the UTF-8 flag in STRING. The data in STRING is
909 B<not> checked for being well-formed UTF-8. Do not use unless you
910 B<know> that the STRING is well-formed UTF-8. Returns the previous
911 state of the UTF-8 flag (so please don't test the return value as
912 I<not> success or failure), or C<undef> if STRING is not a string.
918 [INTERNAL] Turn off the UTF-8 flag in STRING. Do not use frivolously.
919 Returns the previous state of the UTF-8 flag (so please don't test the
920 return value as I<not> success or failure), or C<undef> if STRING is
925 =head1 IMPLEMENTATION CLASSES
927 As mentioned above encodings are (in the current implementation at least)
928 defined by objects. The mapping of encoding name to object is via the
931 The values of the hash can currently be either strings or objects.
932 The string form may go away in the future. The string form occurs
933 when C<encodings()> has scanned C<@INC> for loadable encodings but has
934 not actually loaded the encoding in question. This is because the
935 current "loading" process is all Perl and a bit slow.
937 Once an encoding is loaded then value of the hash is object which
938 implements the encoding. The object should provide the following
945 Should return the string representing the canonical name of the encoding.
947 =item -E<gt>new_sequence
949 This is a placeholder for encodings with state. It should return an
950 object which implements this interface, all current implementations
951 return the original object.
953 =item -E<gt>encode($string,$check)
955 Should return the octet sequence representing I<$string>. If I<$check>
956 is true it should modify I<$string> in place to remove the converted
957 part (i.e. the whole string unless there is an error). If an error
958 occurs it should return the octet sequence for the fragment of string
959 that has been converted, and modify $string in-place to remove the
960 converted part leaving it starting with the problem fragment.
962 If check is is false then C<encode> should make a "best effort" to
963 convert the string - for example by using a replacement character.
965 =item -E<gt>decode($octets,$check)
967 Should return the string that I<$octets> represents. If I<$check> is
968 true it should modify I<$octets> in place to remove the converted part
969 (i.e. the whole sequence unless there is an error). If an error
970 occurs it should return the fragment of string that has been
971 converted, and modify $octets in-place to remove the converted part
972 leaving it starting with the problem fragment.
974 If check is is false then C<decode> should make a "best effort" to
975 convert the string - for example by using Unicode's "\x{FFFD}" as a
976 replacement character.
980 It should be noted that the check behaviour is different from the
981 outer public API. The logic is that the "unchecked" case is useful
982 when encoding is part of a stream which may be reporting errors
983 (e.g. STDERR). In such cases it is desirable to get everything
984 through somehow without causing additional errors which obscure the
985 original one. Also the encoding is best placed to know what the
986 correct replacement character is, so if that is the desired behaviour
987 then letting low level code do it is the most efficient.
989 In contrast if check is true, the scheme above allows the encoding to
990 do as much as it can and tell layer above how much that was. What is
991 lacking at present is a mechanism to report what went wrong. The most
992 likely interface will be an additional method call to the object, or
993 perhaps (to avoid forcing per-stream objects on otherwise stateless
994 encodings) and additional parameter.
996 It is also highly desirable that encoding classes inherit from
997 C<Encode::Encoding> as a base class. This allows that class to define
998 additional behaviour for all encoding objects. For example built in
999 Unicode, UCS-2 and UTF-8 classes use :
1001 package Encode::MyEncoding;
1002 use base qw(Encode::Encoding);
1004 __PACKAGE__->Define(qw(myCanonical myAlias));
1006 To create an object with bless {Name => ...},$class, and call
1007 define_encoding. They inherit their C<name> method from
1008 C<Encode::Encoding>.
1010 =head2 Compiled Encodings
1012 F<Encode.xs> provides a class C<Encode::XS> which provides the
1013 interface described above. It calls a generic octet-sequence to
1014 octet-sequence "engine" that is driven by tables (defined in
1015 F<encengine.c>). The same engine is used for both encode and
1016 decode. C<Encode:XS>'s C<encode> forces Perl's characters to their
1017 UTF-8 form and then treats them as just another multibyte
1018 encoding. C<Encode:XS>'s C<decode> transforms the sequence and then
1019 turns the UTF-8-ness flag as that is the form that the tables are
1020 defined to produce. For details of the engine see the comments in
1023 The tables are produced by the Perl script F<compile> (the name needs
1024 to change so we can eventually install it somewhere). F<compile> can
1025 currently read two formats:
1031 This is a coined format used by Tcl. It is documented in
1032 Encode/EncodeFormat.pod.
1036 This is the semi-standard format used by IBM's ICU package.
1040 F<compile> can write the following forms:
1046 See above - the F<Encode/*.ucm> files provided with the distribution have
1047 been created from the original Tcl .enc files using this approach.
1051 Produces tables as C data structures - this is used to build in encodings
1052 into F<Encode.so>/F<Encode.dll>.
1056 In theory this allows encodings to be stand-alone loadable Perl
1057 extensions. The process has not yet been tested. The plan is to use
1058 this approach for large East Asian encodings.
1062 The set of encodings built-in to F<Encode.so>/F<Encode.dll> is
1063 determined by F<Makefile.PL>. The current set is as follows:
1067 =item ascii and iso-8859-*
1069 That is all the common 8-bit "western" encodings.
1071 =item IBM-1047 and two other variants of EBCDIC.
1073 These are the same variants that are supported by EBCDIC Perl as
1074 "native" encodings. They are included to prove "reversibility" of
1075 some constructs in EBCDIC Perl.
1077 =item symbol and dingbats as used by Tk on X11.
1079 (The reason Encode got started was to support Perl/Tk.)
1083 That set is rather ad hoc and has been driven by the needs of the
1084 tests rather than the needs of typical applications. It is likely
1089 L<perlunicode>, L<perlebcdic>, L<perlfunc/open>, L<PerlIO>, L<encoding>