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 );
61 our %external_tables =
63 'euc-cn' => 'Encode/CN.pm',
64 gb2312 => 'Encode/CN.pm',
65 gb12345 => 'Encode/CN.pm',
66 gbk => 'Encode/CN.pm',
67 cp936 => 'Encode/CN.pm',
68 'iso-ir-165' => 'Encode/CN.pm',
69 'euc-jp' => 'Encode/JP.pm',
70 shiftjis => 'Encode/JP.pm',
71 macjapan => 'Encode/JP.pm',
72 cp932 => 'Encode/JP.pm',
73 'euc-kr' => 'Encode/KR.pm',
74 ksc5601 => 'Encode/KR.pm',
75 cp949 => 'Encode/KR.pm',
76 big5 => 'Encode/TW.pm',
77 'big5-hkscs' => 'Encode/TW.pm',
78 cp950 => 'Encode/TW.pm',
81 eval 'require Encode::HanExtra';
83 push %external_tables,
85 gb18030 => 'Encode/CN.pm', # HanExtra
86 big5plus => 'Encode/TW.pm', # HanExtra
87 'euc-tw', => 'Encode/TW.pm', # HanExtra
97 sort { $a->[1] cmp $b->[1] }
99 grep { $_ ne 'Internal' }
107 # print "# findAlias $_\n";
108 unless (exists $alias{$_})
110 for (my $i=0; $i < @alias; $i += 2)
112 my $alias = $alias[$i];
113 my $val = $alias[$i+1];
115 if (ref($alias) eq 'Regexp' && $_ =~ $alias)
119 elsif (ref($alias) eq 'CODE')
121 $new = &{$alias}($val)
123 elsif (lc($_) eq lc($alias))
129 next if $new eq $_; # avoid (direct) recursion on bugs
130 my $enc = (ref($new)) ? $new : find_encoding($new);
146 my ($alias,$name) = splice(@_,0,2);
147 push(@alias, $alias => $name);
151 # Allow variants of iso-8859-1 etc.
152 define_alias( qr/^iso[-_]?(\d+)[-_](\d+)$/i => '"iso-$1-$2"' );
154 # At least HP-UX has these.
155 define_alias( qr/^iso8859(\d+)$/i => '"iso-8859-$1"' );
158 define_alias( qr/^(?:hp-)?(arabic|greek|hebrew|kana|roman|thai|turkish)8$/i => '"${1}8"' );
160 # The Official name of ASCII.
161 define_alias( qr/^ANSI[-_]?X3\.4[-_]?1968$/i => '"ascii"' );
163 # This is a font issue, not an encoding issue.
164 # (The currency symbol of the Latin 1 upper half
165 # has been redefined as the euro symbol.)
166 define_alias( qr/^(.+)\@euro$/i => '"$1"' );
168 # Allow latin-1 style names as well
169 define_alias( qr/^(?:iso[-_]?)?latin[-_]?(\d+)$/i => '"iso-8859-$latin2iso_num[$1]"' );
171 # Allow winlatin1 style names as well
172 define_alias( qr/^win(latin[12]|cyrillic|baltic|greek|turkish|hebrew|arabic|baltic|vietnamese)$/i => '"cp$winlatin2cp{\u$1}"' );
174 # Common names for non-latin prefered MIME names
175 define_alias( 'ascii' => 'US-ascii',
176 'cyrillic' => 'iso-8859-5',
177 'arabic' => 'iso-8859-6',
178 'greek' => 'iso-8859-7',
179 'hebrew' => 'iso-8859-8',
180 'thai' => 'iso-8859-11',
181 'tis620' => 'iso-8859-11',
184 # At least AIX has IBM-NNN (surprisingly...) instead of cpNNN.
185 # And Microsoft has their own naming (again, surprisingly).
186 define_alias( qr/^(?:ibm|ms)[-_]?(\d\d\d\d?)$/i => '"cp$1"');
188 # Sometimes seen with a leading zero.
189 define_alias( qr/^cp037$/i => '"cp37"');
192 define_alias( qr/^macRomanian$/i => '"macRumanian"');
194 # Standardize on the dashed versions.
195 define_alias( qr/^utf8$/i => 'utf-8' );
196 define_alias( qr/^koi8r$/i => 'koi8-r' );
197 define_alias( qr/^koi8u$/i => 'koi8-u' );
199 # Seen in some Linuxes.
200 define_alias( qr/^ujis$/i => 'euc-jp' );
202 # CP936 doesn't have vendor-addon for GBK, so they're identical.
203 define_alias( qr/^gbk$/i => '"cp936"');
205 # TODO: HP-UX '8' encodings arabic8 greek8 hebrew8 kana8 thai8 turkish8
206 # TODO: HP-UX '15' encodings japanese15 korean15 roi15
207 # TODO: Cyrillic encoding ISO-IR-111 (useful?)
208 # TODO: Armenian encoding ARMSCII-8
209 # TODO: Hebrew encoding ISO-8859-8-1
210 # TODO: Thai encoding TCVN
211 # TODO: Korean encoding Johab
212 # TODO: Vietnamese encodings VPS
213 # TODO: Japanese encoding JIS (not the same as SJIS)
214 # TODO: Mac Asian+African encodings: Arabic Armenian Bengali Burmese
215 # ChineseSimp ChineseTrad Devanagari Ethiopic ExtArabic
216 # Farsi Georgian Gujarati Gurmukhi Hebrew Japanese
217 # Kannada Khmer Korean Laotian Malayalam Mongolian
218 # Oriya Sinhalese Symbol Tamil Telugu Tibetan Vietnamese
220 # Map white space and _ to '-'
221 define_alias( qr/^(\S+)[\s_]+(.*)$/i => '"$1-$2"' );
227 $encoding{$name} = $obj;
229 define_alias($lc => $obj) unless $lc eq $name;
233 define_alias($alias,$obj);
240 my ($class,$name) = @_;
242 if (ref($name) && $name->can('new_sequence'))
247 if (exists $encoding{$name})
249 return $encoding{$name};
251 if (exists $encoding{$lc})
253 return $encoding{$lc};
255 if (exists $external_tables{$lc})
257 require $external_tables{$lc};
258 return $encoding{$name} if exists $encoding{$name};
261 my $oc = $class->findAlias($name);
262 return $oc if defined $oc;
263 return $class->findAlias($lc) if $lc ne $name;
271 return __PACKAGE__->getEncoding($name);
276 my ($name,$string,$check) = @_;
277 my $enc = find_encoding($name);
278 croak("Unknown encoding '$name'") unless defined $enc;
279 my $octets = $enc->encode($string,$check);
280 return undef if ($check && length($string));
286 my ($name,$octets,$check) = @_;
287 my $enc = find_encoding($name);
288 croak("Unknown encoding '$name'") unless defined $enc;
289 my $string = $enc->decode($octets,$check);
290 $_[1] = $octets if $check;
296 my ($string,$from,$to,$check) = @_;
297 my $f = find_encoding($from);
298 croak("Unknown encoding '$from'") unless defined $f;
299 my $t = find_encoding($to);
300 croak("Unknown encoding '$to'") unless defined $t;
301 my $uni = $f->decode($string,$check);
302 return undef if ($check && length($string));
303 $string = $t->encode($uni,$check);
304 return undef if ($check && length($uni));
305 return length($_[0] = $string);
318 return undef unless utf8::decode($str);
322 require Encode::Encoding;
324 require Encode::Internal;
325 require Encode::Unicode;
326 require Encode::utf8;
327 require Encode::iso10646_1;
328 require Encode::ucs2_le;
336 Encode - character encodings
344 The C<Encode> module provides the interfaces between Perl's strings
345 and the rest of the system. Perl strings are sequences of B<characters>.
347 The repertoire of characters that Perl can represent is at least that
348 defined by the Unicode Consortium. On most platforms the ordinal
349 values of the characters (as returned by C<ord(ch)>) is the "Unicode
350 codepoint" for the character (the exceptions are those platforms where
351 the legacy encoding is some variant of EBCDIC rather than a super-set
352 of ASCII - see L<perlebcdic>).
354 Traditionaly computer data has been moved around in 8-bit chunks
355 often called "bytes". These chunks are also known as "octets" in
356 networking standards. Perl is widely used to manipulate data of
357 many types - not only strings of characters representing human or
358 computer languages but also "binary" data being the machines representation
359 of numbers, pixels in an image - or just about anything.
361 When Perl is processing "binary data" the programmer wants Perl to process
362 "sequences of bytes". This is not a problem for Perl - as a byte has 256
363 possible values it easily fits in Perl's much larger "logical character".
365 Due to size concerns, each of B<CJK> (Chinese, Japanese & Korean) modules
366 are not loaded in memory until the first time they're used. Although you
367 don't have to C<use> the corresponding B<Encode::>(B<TW>|B<CN>|B<JP>|B<KR>)
368 modules first, be aware that those encodings will not be in C<%encodings>
369 until their module is loaded (either implicitly through using encodings
370 contained in the same module, or via an explicit C<use>).
378 I<character>: a character in the range 0..(2**32-1) (or more).
379 (What Perl's strings are made of.)
383 I<byte>: a character in the range 0..255
384 (A special case of a Perl character.)
388 I<octet>: 8 bits of data, with ordinal values 0..255
389 (Term for bytes passed to or from a non-Perl context, e.g. disk file.)
393 The marker [INTERNAL] marks Internal Implementation Details, in
394 general meant only for those who think they know what they are doing,
395 and such details may change in future releases.
399 =head2 Characteristics of an Encoding
401 An encoding has a "repertoire" of characters that it can represent,
402 and for each representable character there is at least one sequence of
403 octets that represents it.
405 =head2 Types of Encodings
407 Encodings can be divided into the following types:
411 =item * Fixed length 8-bit (or less) encodings.
413 Each character is a single octet so may have a repertoire of up to
414 256 characters. ASCII and iso-8859-* are typical examples.
416 =item * Fixed length 16-bit encodings
418 Each character is two octets so may have a repertoire of up to
419 65 536 characters. Unicode's UCS-2 is an example. Also used for
420 encodings for East Asian languages.
422 =item * Fixed length 32-bit encodings.
424 Not really very "encoded" encodings. The Unicode code points
425 are just represented as 4-octet integers. None the less because
426 different architectures use different representations of integers
427 (so called "endian") there at least two disctinct encodings.
429 =item * Multi-byte encodings
431 The number of octets needed to represent a character varies.
432 UTF-8 is a particularly complex but regular case of a multi-byte
433 encoding. Several East Asian countries use a multi-byte encoding
434 where 1-octet is used to cover western roman characters and Asian
435 characters get 2-octets.
436 (UTF-16 is strictly a multi-byte encoding taking either 2 or 4 octets
437 to represent a Unicode code point.)
439 =item * "Escape" encodings.
441 These encodings embed "escape sequences" into the octet sequence
442 which describe how the following octets are to be interpreted.
443 The iso-2022-* family is typical. Following the escape sequence
444 octets are encoded by an "embedded" encoding (which will be one
445 of the above types) until another escape sequence switches to
446 a different "embedded" encoding.
448 These schemes are very flexible and can handle mixed languages but are
449 very complex to process (and have state). No escape encodings are
450 implemented for Perl yet.
454 =head2 Specifying Encodings
456 Encodings can be specified to the API described below in two ways:
462 Encoding names are strings with characters taken from a restricted
463 repertoire. See L</"Encoding Names">.
465 =item 2. As an object
467 Encoding objects are returned by C<find_encoding($name)>.
471 =head2 Encoding Names
473 Encoding names are case insensitive. White space in names is ignored.
474 In addition an encoding may have aliases. Each encoding has one
475 "canonical" name. The "canonical" name is chosen from the names of
476 the encoding by picking the first in the following sequence:
480 =item * The MIME name as defined in IETF RFCs.
482 =item * The name in the IANA registry.
484 =item * The name used by the organization that defined it.
488 Because of all the alias issues, and because in the general case
489 encodings have state C<Encode> uses the encoding object internally
490 once an operation is in progress.
492 As of Perl 5.8.0, at least the following encodings are recognized
493 (the => marks aliases):
507 The ISO 8859 and KOI:
509 ISO 8859-1 ISO 8859-6 ISO 8859-11 KOI8-F
510 ISO 8859-2 ISO 8859-7 (12 doesn't exist) KOI8-R
511 ISO 8859-3 ISO 8859-8 ISO 8859-13 KOI8-U
512 ISO 8859-4 ISO 8859-9 ISO 8859-14
513 ISO 8859-5 ISO 8859-10 ISO 8859-15
516 Latin1 => 8859-1 Latin6 => 8859-10
517 Latin2 => 8859-2 Latin7 => 8859-13
518 Latin3 => 8859-3 Latin8 => 8859-14
519 Latin4 => 8859-4 Latin9 => 8859-15
520 Latin5 => 8859-9 Latin10 => 8859-16
529 The CJKV: Chinese, Japanese, Korean, Vietnamese:
531 ISO 2022 ISO 2022 JP-1 JIS 0201 GB 1988 Big5 EUC-CN
532 ISO 2022 CN ISO 2022 JP-2 JIS 0208 GB 2312 HZ EUC-JP
533 ISO 2022 JP ISO 2022 KR JIS 0210 GB 12345 CNS 11643 EUC-JP-0212
534 Shift-JIS GBK Big5-HKSCS EUC-KR
537 (Due to size concerns, additional Chinese encodings including C<GB 18030>,
538 C<EUC-TW> and C<BIG5PLUS> are distributed separately on CPAN, under the name
539 L<Encode::HanExtra>.)
543 CP37 CP852 CP861 CP866 CP949 CP1251 CP1256
544 CP424 CP855 CP862 CP869 CP950 CP1252 CP1257
545 CP737 CP856 CP863 CP874 CP1006 CP1253 CP1258
546 CP775 CP857 CP864 CP932 CP1047 CP1254
547 CP850 CP860 CP865 CP936 CP1250 CP1255
551 WinCyrillic => CP1251
553 WinTurkiskh => CP1254
557 WinVietnamese => CP1258
559 (All the CPI<NNN...> are available also as IBMI<NNN...>.)
563 MacCentralEuropean MacJapanese
565 MacCyrillic MacRomanian
568 MacIcelandic MacTurkish
579 =head1 PERL ENCODING API
581 =head2 Generic Encoding Interface
587 $bytes = encode(ENCODING, $string[, CHECK])
589 Encodes string from Perl's internal form into I<ENCODING> and returns
590 a sequence of octets. For CHECK see L</"Handling Malformed Data">.
592 For example to convert (internally UTF-8 encoded) Unicode data
595 $octets = encode("utf8", $unicode);
599 $string = decode(ENCODING, $bytes[, CHECK])
601 Decode sequence of octets assumed to be in I<ENCODING> into Perl's
602 internal form and returns the resulting string. For CHECK see
603 L</"Handling Malformed Data">.
605 For example to convert ISO 8859-1 data to UTF-8:
607 $utf8 = decode("latin1", $latin1);
611 from_to($string, FROM_ENCODING, TO_ENCODING[, CHECK])
613 Convert B<in-place> the data between two encodings. How did the data
614 in $string originally get to be in FROM_ENCODING? Either using
615 encode() or through PerlIO: See L</"Encoding and IO">. For CHECK
616 see L</"Handling Malformed Data">.
618 For example to convert ISO 8859-1 data to UTF-8:
620 from_to($data, "iso-8859-1", "utf-8");
622 and to convert it back:
624 from_to($data, "utf-8", "iso-8859-1");
626 Note that because the conversion happens in place, the data to be
627 converted cannot be a string constant, it must be a scalar variable.
631 =head2 Handling Malformed Data
633 If CHECK is not set, C<undef> is returned. If the data is supposed to
634 be UTF-8, an optional lexical warning (category utf8) is given. If
635 CHECK is true but not a code reference, dies.
637 It would desirable to have a way to indicate that transform should use
638 the encodings "replacement character" - no such mechanism is defined yet.
640 It is also planned to allow I<CHECK> to be a code reference.
642 This is not yet implemented as there are design issues with what its
643 arguments should be and how it returns its results.
649 Passed remaining fragment of string being processed.
650 Modifies it in place to remove bytes/characters it can understand
651 and returns a string used to represent them.
655 my $ch = substr($_[0],0,1,'');
656 return sprintf("\x{%02X}",ord($ch);
659 This scheme is close to how underlying C code for Encode works, but gives
660 the fixup routine very little context.
664 Passed original string, and an index into it of the problem area, and
665 output string so far. Appends what it will to output string and
666 returns new index into original string. For example:
669 # my ($s,$i,$d) = @_;
670 my $ch = substr($_[0],$_[1],1);
671 $_[2] .= sprintf("\x{%02X}",ord($ch);
675 This scheme gives maximal control to the fixup routine but is more
676 complicated to code, and may need internals of Encode to be tweaked to
677 keep original string intact.
683 Multiple return values rather than in-place modifications.
685 Index into the string could be pos($str) allowing s/\G...//.
691 The Unicode consortium defines the UTF-8 standard as a way of encoding
692 the entire Unicode repertiore as sequences of octets. This encoding is
693 expected to become very widespread. Perl can use this form internaly
694 to represent strings, so conversions to and from this form are
695 particularly efficient (as octets in memory do not have to change,
696 just the meta-data that tells Perl how to treat them).
702 $bytes = encode_utf8($string);
704 The characters that comprise string are encoded in Perl's superset of UTF-8
705 and the resulting octets returned as a sequence of bytes. All possible
706 characters have a UTF-8 representation so this function cannot fail.
710 $string = decode_utf8($bytes [,CHECK]);
712 The sequence of octets represented by $bytes is decoded from UTF-8
713 into a sequence of logical characters. Not all sequences of octets
714 form valid UTF-8 encodings, so it is possible for this call to fail.
715 For CHECK see L</"Handling Malformed Data">.
719 =head2 Other Encodings of Unicode
721 UTF-16 is similar to UCS-2, 16 bit or 2-byte chunks. UCS-2 can only
722 represent 0..0xFFFF, while UTF-16 has a I<surrogate pair> scheme which
723 allows it to cover the whole Unicode range.
725 Surrogates are code points set aside to encode the 0x01000..0x10FFFF
726 range of Unicode code points in pairs of 16-bit units. The I<high
727 surrogates> are the range 0xD800..0xDBFF, and the I<low surrogates>
728 are the range 0xDC00..0xDFFFF. The surrogate encoding is
730 $hi = ($uni - 0x10000) / 0x400 + 0xD800;
731 $lo = ($uni - 0x10000) % 0x400 + 0xDC00;
735 $uni = 0x10000 + ($hi - 0xD8000) * 0x400 + ($lo - 0xDC00);
737 Encode implements big-endian UCS-2 aliased to "iso-10646-1" as that
738 happens to be the name used by that representation when used with X11
741 UTF-32 or UCS-4 is 32-bit or 4-byte chunks. Perl's logical characters
742 can be considered as being in this form without encoding. An encoding
743 to transfer strings in this form (e.g. to write them to a file) would
746 pack('L*', unpack('U*', $string)); # native
748 pack('V*', unpack('U*', $string)); # little-endian
750 pack('N*', unpack('U*', $string)); # big-endian
752 depending on the endianness required.
754 No UTF-32 encodings are implemented yet.
756 Both UCS-2 and UCS-4 style encodings can have "byte order marks" by
757 representing the code point 0xFFFE as the very first thing in a file.
759 =head2 Listing available encodings
761 use Encode qw(encodings);
764 Returns a list of the canonical names of the available encodings.
766 =head2 Defining Aliases
768 use Encode qw(define_alias);
769 define_alias( newName => ENCODING);
771 Allows newName to be used as am alias for ENCODING. ENCODING may be
772 either the name of an encoding or and encoding object (as above).
774 Currently I<newName> can be specified in the following ways:
778 =item As a simple string.
780 =item As a qr// compiled regular expression, e.g.:
782 define_alias( qr/^iso8859-(\d+)$/i => '"iso-8859-$1"' );
784 In this case if I<ENCODING> is not a reference it is C<eval>-ed to
785 allow C<$1> etc. to be subsituted. The example is one way to names as
786 used in X11 font names to alias the MIME names for the iso-8859-*
789 =item As a code reference, e.g.:
791 define_alias( sub { return /^iso8859-(\d+)$/i ? "iso-8859-$1" : undef } , '');
793 In this case C<$_> will be set to the name that is being looked up and
794 I<ENCODING> is passed to the sub as its first argument. The example
795 is another way to names as used in X11 font names to alias the MIME
796 names for the iso-8859-* family.
800 =head2 Defining Encodings
802 use Encode qw(define_alias);
803 define_encoding( $object, 'canonicalName' [,alias...]);
805 Causes I<canonicalName> to be associated with I<$object>. The object
806 should provide the interface described in L</"IMPLEMENTATION CLASSES">
807 below. If more than two arguments are provided then additional
808 arguments are taken as aliases for I<$object> as for C<define_alias>.
810 =head1 Encoding and IO
812 It is very common to want to do encoding transformations when
813 reading or writing files, network connections, pipes etc.
814 If Perl is configured to use the new 'perlio' IO system then
815 C<Encode> provides a "layer" (See L<perliol>) which can transform
816 data as it is read or written.
818 Here is how the blind poet would modernise the encoding:
821 open(my $iliad,'<:encoding(iso-8859-7)','iliad.greek');
822 open(my $utf8,'>:utf8','iliad.utf8');
828 In addition the new IO system can also be configured to read/write
829 UTF-8 encoded characters (as noted above this is efficient):
831 open(my $fh,'>:utf8','anything');
832 print $fh "Any \x{0021} string \N{SMILEY FACE}\n";
834 Either of the above forms of "layer" specifications can be made the default
835 for a lexical scope with the C<use open ...> pragma. See L<open>.
837 Once a handle is open is layers can be altered using C<binmode>.
839 Without any such configuration, or if Perl itself is built using
840 system's own IO, then write operations assume that file handle accepts
841 only I<bytes> and will C<die> if a character larger than 255 is
842 written to the handle. When reading, each octet from the handle
843 becomes a byte-in-a-character. Note that this default is the same
844 behaviour as bytes-only languages (including Perl before v5.6) would
845 have, and is sufficient to handle native 8-bit encodings
846 e.g. iso-8859-1, EBCDIC etc. and any legacy mechanisms for handling
847 other encodings and binary data.
849 In other cases it is the programs responsibility to transform
850 characters into bytes using the API above before doing writes, and to
851 transform the bytes read from a handle into characters before doing
852 "character operations" (e.g. C<lc>, C</\W+/>, ...).
854 You can also use PerlIO to convert larger amounts of data you don't
855 want to bring into memory. For example to convert between ISO 8859-1
856 (Latin 1) and UTF-8 (or UTF-EBCDIC in EBCDIC machines):
858 open(F, "<:encoding(iso-8859-1)", "data.txt") or die $!;
859 open(G, ">:utf8", "data.utf") or die $!;
860 while (<F>) { print G }
862 # Could also do "print G <F>" but that would pull
863 # the whole file into memory just to write it out again.
867 open(my $f, "<:encoding(cp1252)")
868 open(my $g, ">:encoding(iso-8859-2)")
869 open(my $h, ">:encoding(latin9)") # iso-8859-15
871 See L<PerlIO> for more information.
873 See also L<encoding> for how to change the default encoding of the
876 =head1 Encoding How to ...
882 =item * IO with mixed content (faking iso-2020-*)
884 =item * MIME's Content-Length:
886 =item * UTF-8 strings in binary data.
888 =item * Perl/Encode wrappers on non-Unicode XS modules.
892 =head1 Messing with Perl's Internals
894 The following API uses parts of Perl's internals in the current
895 implementation. As such they are efficient, but may change.
899 =item * is_utf8(STRING [, CHECK])
901 [INTERNAL] Test whether the UTF-8 flag is turned on in the STRING.
902 If CHECK is true, also checks the data in STRING for being well-formed
903 UTF-8. Returns true if successful, false otherwise.
905 =item * valid_utf8(STRING)
907 [INTERNAL] Test whether STRING is in a consistent state. Will return
908 true if string is held as bytes, or is well-formed UTF-8 and has the
909 UTF-8 flag on. Main reason for this routine is to allow Perl's
910 testsuite to check that operations have left strings in a consistent
917 [INTERNAL] Turn on the UTF-8 flag in STRING. The data in STRING is
918 B<not> checked for being well-formed UTF-8. Do not use unless you
919 B<know> that the STRING is well-formed UTF-8. Returns the previous
920 state of the UTF-8 flag (so please don't test the return value as
921 I<not> success or failure), or C<undef> if STRING is not a string.
927 [INTERNAL] Turn off the UTF-8 flag in STRING. Do not use frivolously.
928 Returns the previous state of the UTF-8 flag (so please don't test the
929 return value as I<not> success or failure), or C<undef> if STRING is
934 =head1 IMPLEMENTATION CLASSES
936 As mentioned above encodings are (in the current implementation at least)
937 defined by objects. The mapping of encoding name to object is via the
940 The values of the hash can currently be either strings or objects.
941 The string form may go away in the future. The string form occurs
942 when C<encodings()> has scanned C<@INC> for loadable encodings but has
943 not actually loaded the encoding in question. This is because the
944 current "loading" process is all Perl and a bit slow.
946 Once an encoding is loaded then value of the hash is object which
947 implements the encoding. The object should provide the following
954 Should return the string representing the canonical name of the encoding.
956 =item -E<gt>new_sequence
958 This is a placeholder for encodings with state. It should return an
959 object which implements this interface, all current implementations
960 return the original object.
962 =item -E<gt>encode($string,$check)
964 Should return the octet sequence representing I<$string>. If I<$check>
965 is true it should modify I<$string> in place to remove the converted
966 part (i.e. the whole string unless there is an error). If an error
967 occurs it should return the octet sequence for the fragment of string
968 that has been converted, and modify $string in-place to remove the
969 converted part leaving it starting with the problem fragment.
971 If check is is false then C<encode> should make a "best effort" to
972 convert the string - for example by using a replacement character.
974 =item -E<gt>decode($octets,$check)
976 Should return the string that I<$octets> represents. If I<$check> is
977 true it should modify I<$octets> in place to remove the converted part
978 (i.e. the whole sequence unless there is an error). If an error
979 occurs it should return the fragment of string that has been
980 converted, and modify $octets in-place to remove the converted part
981 leaving it starting with the problem fragment.
983 If check is is false then C<decode> should make a "best effort" to
984 convert the string - for example by using Unicode's "\x{FFFD}" as a
985 replacement character.
989 It should be noted that the check behaviour is different from the
990 outer public API. The logic is that the "unchecked" case is useful
991 when encoding is part of a stream which may be reporting errors
992 (e.g. STDERR). In such cases it is desirable to get everything
993 through somehow without causing additional errors which obscure the
994 original one. Also the encoding is best placed to know what the
995 correct replacement character is, so if that is the desired behaviour
996 then letting low level code do it is the most efficient.
998 In contrast if check is true, the scheme above allows the encoding to
999 do as much as it can and tell layer above how much that was. What is
1000 lacking at present is a mechanism to report what went wrong. The most
1001 likely interface will be an additional method call to the object, or
1002 perhaps (to avoid forcing per-stream objects on otherwise stateless
1003 encodings) and additional parameter.
1005 It is also highly desirable that encoding classes inherit from
1006 C<Encode::Encoding> as a base class. This allows that class to define
1007 additional behaviour for all encoding objects. For example built in
1008 Unicode, UCS-2 and UTF-8 classes use :
1010 package Encode::MyEncoding;
1011 use base qw(Encode::Encoding);
1013 __PACKAGE__->Define(qw(myCanonical myAlias));
1015 To create an object with bless {Name => ...},$class, and call
1016 define_encoding. They inherit their C<name> method from
1017 C<Encode::Encoding>.
1019 =head2 Compiled Encodings
1021 F<Encode.xs> provides a class C<Encode::XS> which provides the
1022 interface described above. It calls a generic octet-sequence to
1023 octet-sequence "engine" that is driven by tables (defined in
1024 F<encengine.c>). The same engine is used for both encode and
1025 decode. C<Encode:XS>'s C<encode> forces Perl's characters to their
1026 UTF-8 form and then treats them as just another multibyte
1027 encoding. C<Encode:XS>'s C<decode> transforms the sequence and then
1028 turns the UTF-8-ness flag as that is the form that the tables are
1029 defined to produce. For details of the engine see the comments in
1032 The tables are produced by the Perl script F<compile> (the name needs
1033 to change so we can eventually install it somewhere). F<compile> can
1034 currently read two formats:
1040 This is a coined format used by Tcl. It is documented in
1041 Encode/EncodeFormat.pod.
1045 This is the semi-standard format used by IBM's ICU package.
1049 F<compile> can write the following forms:
1055 See above - the F<Encode/*.ucm> files provided with the distribution have
1056 been created from the original Tcl .enc files using this approach.
1060 Produces tables as C data structures - this is used to build in encodings
1061 into F<Encode.so>/F<Encode.dll>.
1065 In theory this allows encodings to be stand-alone loadable Perl
1066 extensions. The process has not yet been tested. The plan is to use
1067 this approach for large East Asian encodings.
1071 The set of encodings built-in to F<Encode.so>/F<Encode.dll> is
1072 determined by F<Makefile.PL>. The current set is as follows:
1076 =item ascii and iso-8859-*
1078 That is all the common 8-bit "western" encodings.
1080 =item IBM-1047 and two other variants of EBCDIC.
1082 These are the same variants that are supported by EBCDIC Perl as
1083 "native" encodings. They are included to prove "reversibility" of
1084 some constructs in EBCDIC Perl.
1086 =item symbol and dingbats as used by Tk on X11.
1088 (The reason Encode got started was to support Perl/Tk.)
1092 That set is rather ad hoc and has been driven by the needs of the
1093 tests rather than the needs of typical applications. It is likely
1098 L<perlunicode>, L<perlebcdic>, L<perlfunc/open>, L<PerlIO>, L<encoding>