9 our @ISA = qw(Exporter DynaLoader);
11 # Public, encouraged API is exported by default
37 # Documentation moved after __END__ for speed - NI-S
41 # Make a %encoding package variable to allow a certain amount of cheating
43 my @alias; # ordered matching list
44 my %alias; # cached known aliases
46 # 0 1 2 3 4 5 6 7 8 9 10
47 our @latin2iso_num = ( 0, 1, 2, 3, 4, 9, 10, 13, 14, 15, 16 );
65 return keys %encoding;
72 # print "# findAlias $_\n";
73 unless (exists $alias{$_})
75 for (my $i=0; $i < @alias; $i += 2)
77 my $alias = $alias[$i];
78 my $val = $alias[$i+1];
80 if (ref($alias) eq 'Regexp' && $_ =~ $alias)
84 elsif (ref($alias) eq 'CODE')
86 $new = &{$alias}($val)
88 elsif (lc($_) eq lc($alias))
94 next if $new eq $_; # avoid (direct) recursion on bugs
95 my $enc = (ref($new)) ? $new : find_encoding($new);
111 my ($alias,$name) = splice(@_,0,2);
112 push(@alias, $alias => $name);
116 # Allow variants of iso-8859-1 etc.
117 define_alias( qr/^iso[-_]?(\d+)[-_](\d+)$/i => '"iso-$1-$2"' );
119 # At least HP-UX has these.
120 define_alias( qr/^iso8859(\d+)$/i => '"iso-8859-$1"' );
123 define_alias( qr/^(?:hp-)?(arabic|greek|hebrew|kana|roman|thai|turkish)8$/i => '"${1}8"' );
125 # The Official name of ASCII.
126 define_alias( qr/^ANSI[-_]?X3\.4[-_]?1968$/i => '"ascii"' );
128 # This is a font issue, not an encoding issue.
129 # (The currency symbol of the Latin 1 upper half
130 # has been redefined as the euro symbol.)
131 define_alias( qr/^(.+)\@euro$/i => '"$1"' );
133 # Allow latin-1 style names as well
134 define_alias( qr/^(?:iso[-_]?)?latin[-_]?(\d+)$/i => '"iso-8859-$latin2iso_num[$1]"' );
136 # Allow winlatin1 style names as well
137 define_alias( qr/^win(latin[12]|cyrillic|baltic|greek|turkish|hebrew|arabic|baltic|vietnamese)$/i => '"cp$winlatin2cp{\u$1}"' );
139 # Common names for non-latin prefered MIME names
140 define_alias( 'ascii' => 'US-ascii',
141 'cyrillic' => 'iso-8859-5',
142 'arabic' => 'iso-8859-6',
143 'greek' => 'iso-8859-7',
144 'hebrew' => 'iso-8859-8',
145 'thai' => 'iso-8859-11',
146 'tis620' => 'iso-8859-11',
149 # At least AIX has IBM-NNN (surprisingly...) instead of cpNNN.
150 define_alias( qr/^ibm[-_]?(\d\d\d\d?)$/i => '"cp$1"');
152 # Standardize on the dashed versions.
153 define_alias( qr/^utf8$/i => 'utf-8' );
154 define_alias( qr/^koi8r$/i => 'koi8-r' );
155 define_alias( qr/^koi8u$/i => 'koi8-u' );
157 # TODO: HP-UX '8' encodings arabic8 greek8 hebrew8 kana8 thai8 turkish8
158 # TODO: HP-UX '15' encodings japanese15 korean15 roi15
159 # TODO: Cyrillic encoding ISO-IR-111 (useful?)
160 # TODO: Chinese encodings GB18030 GBK Big5-HSKCS EUC-TW
161 # TODO: Armenian encoding ARMSCII-8
162 # TODO: Hebrew encoding ISO-8859-8-1
163 # TODO: Thai encoding TCVN
164 # TODO: Korean encoding Johab
165 # TODO: Vietnamese encodings VISCII VPS
166 # TODO: Japanese encoding JIS (not the same as SJIS)
167 # TODO: Mac Asian+African encodings: Arabic Armenian Bengali Burmese
168 # ChineseSimp ChineseTrad Devanagari Ethiopic ExtArabic
169 # Farsi Georgian Gujarati Gurmukhi Hebrew Japanese
170 # Kannada Khmer Korean Laotian Malayalam Mongolian
171 # Oriya Sinhalese Symbol Tamil Telugu Tibetan Vietnamese
172 # TODO: what is the Japanese 'UJIS' encoding seen in some Linuxes?
174 # Map white space and _ to '-'
175 define_alias( qr/^(\S+)[\s_]+(.*)$/i => '"$1-$2"' );
181 $encoding{$name} = $obj;
183 define_alias($lc => $obj) unless $lc eq $name;
187 define_alias($alias,$obj);
194 my ($class,$name) = @_;
196 if (ref($name) && $name->can('new_sequence'))
201 if (exists $encoding{$name})
203 return $encoding{$name};
205 if (exists $encoding{$lc})
207 return $encoding{$lc};
210 my $oc = $class->findAlias($name);
211 return $oc if defined $oc;
212 return $class->findAlias($lc) if $lc ne $name;
220 return __PACKAGE__->getEncoding($name);
225 my ($name,$string,$check) = @_;
226 my $enc = find_encoding($name);
227 croak("Unknown encoding '$name'") unless defined $enc;
228 my $octets = $enc->encode($string,$check);
229 return undef if ($check && length($string));
235 my ($name,$octets,$check) = @_;
236 my $enc = find_encoding($name);
237 croak("Unknown encoding '$name'") unless defined $enc;
238 my $string = $enc->decode($octets,$check);
239 $_[1] = $octets if $check;
245 my ($string,$from,$to,$check) = @_;
246 my $f = find_encoding($from);
247 croak("Unknown encoding '$from'") unless defined $f;
248 my $t = find_encoding($to);
249 croak("Unknown encoding '$to'") unless defined $t;
250 my $uni = $f->decode($string,$check);
251 return undef if ($check && length($string));
252 $string = $t->encode($uni,$check);
253 return undef if ($check && length($uni));
254 return length($_[0] = $string);
267 return pack("C*", unpack("U*", $str)) if is_utf8($str);
268 return undef unless utf8::decode($str);
272 package Encode::Encoding;
273 # Base class for classes which implement encodings
278 my $canonical = shift;
279 $obj = bless { Name => $canonical },$obj unless ref $obj;
280 # warn "$canonical => $obj\n";
281 Encode::define_encoding($obj, $canonical, @_);
284 sub name { shift->{'Name'} }
286 # Temporary legacy methods
287 sub toUnicode { shift->decode(@_) }
288 sub fromUnicode { shift->encode(@_) }
290 sub new_sequence { return $_[0] }
293 use base 'Encode::Encoding';
295 package Encode::Internal;
296 use base 'Encode::Encoding';
298 # Dummy package that provides the encode interface but leaves data
299 # as UTF-X encoded. It is here so that from_to() works.
301 __PACKAGE__->Define('Internal');
303 Encode::define_alias( 'Unicode' => 'Internal' ) if ord('A') == 65;
307 my ($obj,$str,$chk) = @_;
315 package Encoding::Unicode;
316 use base 'Encode::Encoding';
318 __PACKAGE__->Define('Unicode') unless ord('A') == 65;
322 my ($obj,$str,$chk) = @_;
324 for (my $i = 0; $i < length($str); $i++)
326 $res .= chr(utf8::unicode_to_native(ord(substr($str,$i,1))));
334 my ($obj,$str,$chk) = @_;
336 for (my $i = 0; $i < length($str); $i++)
338 $res .= chr(utf8::native_to_unicode(ord(substr($str,$i,1))));
345 package Encode::utf8;
346 use base 'Encode::Encoding';
347 # package to allow long-hand
348 # $octets = encode( utf8 => $string );
351 __PACKAGE__->Define(qw(UTF-8 utf8));
355 my ($obj,$octets,$chk) = @_;
356 my $str = Encode::decode_utf8($octets);
367 my ($obj,$string,$chk) = @_;
368 my $octets = Encode::encode_utf8($string);
373 package Encode::iso10646_1;
374 use base 'Encode::Encoding';
375 # Encoding is 16-bit network order Unicode (no surogates)
376 # Used for X font encodings
378 __PACKAGE__->Define(qw(UCS-2 iso-10646-1));
382 my ($obj,$str,$chk) = @_;
386 my $code = unpack('n',substr($str,0,2,'')) & 0xffff;
389 $_[1] = $str if $chk;
396 my ($obj,$uni,$chk) = @_;
400 my $ch = substr($uni,0,1,'');
407 $str .= pack('n',$x);
409 $_[1] = $uni if $chk;
413 package Encode::ucs_2le;
414 use base 'Encode::Encoding';
416 __PACKAGE__->Define(qw(UCS-2le UCS-2LE ucs-2le));
420 my ($obj,$str,$chk) = @_;
424 my $code = unpack('v',substr($str,0,2,'')) & 0xffff;
427 $_[1] = $str if $chk;
434 my ($obj,$uni,$chk) = @_;
438 my $ch = substr($uni,0,1,'');
445 $str .= pack('v',$x);
447 $_[1] = $uni if $chk;
451 # switch back to Encode package in case we ever add AutoLoader
460 Encode - character encodings
468 The C<Encode> module provides the interfaces between Perl's strings
469 and the rest of the system. Perl strings are sequences of B<characters>.
471 The repertoire of characters that Perl can represent is at least that
472 defined by the Unicode Consortium. On most platforms the ordinal
473 values of the characters (as returned by C<ord(ch)>) is the "Unicode
474 codepoint" for the character (the exceptions are those platforms where
475 the legacy encoding is some variant of EBCDIC rather than a super-set
476 of ASCII - see L<perlebcdic>).
478 Traditionaly computer data has been moved around in 8-bit chunks
479 often called "bytes". These chunks are also known as "octets" in
480 networking standards. Perl is widely used to manipulate data of
481 many types - not only strings of characters representing human or
482 computer languages but also "binary" data being the machines representation
483 of numbers, pixels in an image - or just about anything.
485 When Perl is processing "binary data" the programmer wants Perl to process
486 "sequences of bytes". This is not a problem for Perl - as a byte has 256
487 possible values it easily fits in Perl's much larger "logical character".
495 I<character>: a character in the range 0..(2**32-1) (or more).
496 (What Perl's strings are made of.)
500 I<byte>: a character in the range 0..255
501 (A special case of a Perl character.)
505 I<octet>: 8 bits of data, with ordinal values 0..255
506 (Term for bytes passed to or from a non-Perl context, e.g. disk file.)
510 The marker [INTERNAL] marks Internal Implementation Details, in
511 general meant only for those who think they know what they are doing,
512 and such details may change in future releases.
516 =head2 Characteristics of an Encoding
518 An encoding has a "repertoire" of characters that it can represent,
519 and for each representable character there is at least one sequence of
520 octets that represents it.
522 =head2 Types of Encodings
524 Encodings can be divided into the following types:
528 =item * Fixed length 8-bit (or less) encodings.
530 Each character is a single octet so may have a repertoire of up to
531 256 characters. ASCII and iso-8859-* are typical examples.
533 =item * Fixed length 16-bit encodings
535 Each character is two octets so may have a repertoire of up to
536 65 536 characters. Unicode's UCS-2 is an example. Also used for
537 encodings for East Asian languages.
539 =item * Fixed length 32-bit encodings.
541 Not really very "encoded" encodings. The Unicode code points
542 are just represented as 4-octet integers. None the less because
543 different architectures use different representations of integers
544 (so called "endian") there at least two disctinct encodings.
546 =item * Multi-byte encodings
548 The number of octets needed to represent a character varies.
549 UTF-8 is a particularly complex but regular case of a multi-byte
550 encoding. Several East Asian countries use a multi-byte encoding
551 where 1-octet is used to cover western roman characters and Asian
552 characters get 2-octets.
553 (UTF-16 is strictly a multi-byte encoding taking either 2 or 4 octets
554 to represent a Unicode code point.)
556 =item * "Escape" encodings.
558 These encodings embed "escape sequences" into the octet sequence
559 which describe how the following octets are to be interpreted.
560 The iso-2022-* family is typical. Following the escape sequence
561 octets are encoded by an "embedded" encoding (which will be one
562 of the above types) until another escape sequence switches to
563 a different "embedded" encoding.
565 These schemes are very flexible and can handle mixed languages but are
566 very complex to process (and have state). No escape encodings are
567 implemented for Perl yet.
571 =head2 Specifying Encodings
573 Encodings can be specified to the API described below in two ways:
579 Encoding names are strings with characters taken from a restricted
580 repertoire. See L</"Encoding Names">.
582 =item 2. As an object
584 Encoding objects are returned by C<find_encoding($name)>.
588 =head2 Encoding Names
590 Encoding names are case insensitive. White space in names is ignored.
591 In addition an encoding may have aliases. Each encoding has one
592 "canonical" name. The "canonical" name is chosen from the names of
593 the encoding by picking the first in the following sequence:
597 =item * The MIME name as defined in IETF RFC-XXXX.
599 =item * The name in the IANA registry.
601 =item * The name used by the the organization that defined it.
605 Because of all the alias issues, and because in the general case
606 encodings have state C<Encode> uses the encoding object internally
607 once an operation is in progress.
609 =head1 PERL ENCODING API
611 =head2 Generic Encoding Interface
617 $bytes = encode(ENCODING, $string[, CHECK])
619 Encodes string from Perl's internal form into I<ENCODING> and returns
620 a sequence of octets. For CHECK see L</"Handling Malformed Data">.
624 $string = decode(ENCODING, $bytes[, CHECK])
626 Decode sequence of octets assumed to be in I<ENCODING> into Perl's
627 internal form and returns the resulting string. For CHECK see
628 L</"Handling Malformed Data">.
632 from_to($string, FROM_ENCODING, TO_ENCODING[, CHECK])
634 Convert B<in-place> the data between two encodings. How did the data
635 in $string originally get to be in FROM_ENCODING? Either using
636 encode() or through PerlIO: See L</"Encoding and IO">. For CHECK
637 see L</"Handling Malformed Data">.
639 For example to convert ISO 8859-1 data to UTF-8:
641 from_to($data, "iso-8859-1", "utf-8");
643 and to convert it back:
645 from_to($data, "utf-8", "iso-8859-1");
647 Note that because the conversion happens in place, the data to be
648 converted cannot be a string constant, it must be a scalar variable.
652 =head2 Handling Malformed Data
654 If CHECK is not set, C<undef> is returned. If the data is supposed to
655 be UTF-8, an optional lexical warning (category utf8) is given. If
656 CHECK is true but not a code reference, dies.
658 It would desirable to have a way to indicate that transform should use
659 the encodings "replacement character" - no such mechanism is defined yet.
661 It is also planned to allow I<CHECK> to be a code reference.
663 This is not yet implemented as there are design issues with what its
664 arguments should be and how it returns its results.
670 Passed remaining fragment of string being processed.
671 Modifies it in place to remove bytes/characters it can understand
672 and returns a string used to represent them.
676 my $ch = substr($_[0],0,1,'');
677 return sprintf("\x{%02X}",ord($ch);
680 This scheme is close to how underlying C code for Encode works, but gives
681 the fixup routine very little context.
685 Passed original string, and an index into it of the problem area, and
686 output string so far. Appends what it will to output string and
687 returns new index into original string. For example:
690 # my ($s,$i,$d) = @_;
691 my $ch = substr($_[0],$_[1],1);
692 $_[2] .= sprintf("\x{%02X}",ord($ch);
696 This scheme gives maximal control to the fixup routine but is more
697 complicated to code, and may need internals of Encode to be tweaked to
698 keep original string intact.
704 Multiple return values rather than in-place modifications.
706 Index into the string could be pos($str) allowing s/\G...//.
712 The Unicode consortium defines the UTF-8 standard as a way of encoding
713 the entire Unicode repertiore as sequences of octets. This encoding is
714 expected to become very widespread. Perl can use this form internaly
715 to represent strings, so conversions to and from this form are
716 particularly efficient (as octets in memory do not have to change,
717 just the meta-data that tells Perl how to treat them).
723 $bytes = encode_utf8($string);
725 The characters that comprise string are encoded in Perl's superset of UTF-8
726 and the resulting octets returned as a sequence of bytes. All possible
727 characters have a UTF-8 representation so this function cannot fail.
731 $string = decode_utf8($bytes [,CHECK]);
733 The sequence of octets represented by $bytes is decoded from UTF-8
734 into a sequence of logical characters. Not all sequences of octets
735 form valid UTF-8 encodings, so it is possible for this call to fail.
736 For CHECK see L</"Handling Malformed Data">.
740 =head2 Other Encodings of Unicode
742 UTF-16 is similar to UCS-2, 16 bit or 2-byte chunks. UCS-2 can only
743 represent 0..0xFFFF, while UTF-16 has a I<surrogate pair> scheme which
744 allows it to cover the whole Unicode range.
746 Surrogates are code points set aside to encode the 0x01000..0x10FFFF
747 range of Unicode code points in pairs of 16-bit units. The I<high
748 surrogates> are the range 0xD800..0xDBFF, and the I<low surrogates>
749 are the range 0xDC00..0xDFFFF. The surrogate encoding is
751 $hi = ($uni - 0x10000) / 0x400 + 0xD800;
752 $lo = ($uni - 0x10000) % 0x400 + 0xDC00;
756 $uni = 0x10000 + ($hi - 0xD8000) * 0x400 + ($lo - 0xDC00);
758 Encode implements big-endian UCS-2 aliased to "iso-10646-1" as that
759 happens to be the name used by that representation when used with X11
762 UTF-32 or UCS-4 is 32-bit or 4-byte chunks. Perl's logical characters
763 can be considered as being in this form without encoding. An encoding
764 to transfer strings in this form (e.g. to write them to a file) would
767 pack('L*', unpack('U*', $string)); # native
769 pack('V*', unpack('U*', $string)); # little-endian
771 pack('N*', unpack('U*', $string)); # big-endian
773 depending on the endianness required.
775 No UTF-32 encodings are implemented yet.
777 Both UCS-2 and UCS-4 style encodings can have "byte order marks" by
778 representing the code point 0xFFFE as the very first thing in a file.
780 =head2 Listing available encodings
782 use Encode qw(encodings);
785 Returns a list of the canonical names of the available encodings.
787 =head2 Defining Aliases
789 use Encode qw(define_alias);
790 define_alias( newName => ENCODING);
792 Allows newName to be used as am alias for ENCODING. ENCODING may be
793 either the name of an encoding or and encoding object (as above).
795 Currently I<newName> can be specified in the following ways:
799 =item As a simple string.
801 =item As a qr// compiled regular expression, e.g.:
803 define_alias( qr/^iso8859-(\d+)$/i => '"iso-8859-$1"' );
805 In this case if I<ENCODING> is not a reference it is C<eval>-ed to
806 allow C<$1> etc. to be subsituted. The example is one way to names as
807 used in X11 font names to alias the MIME names for the iso-8859-*
810 =item As a code reference, e.g.:
812 define_alias( sub { return /^iso8859-(\d+)$/i ? "iso-8859-$1" : undef } , '');
814 In this case C<$_> will be set to the name that is being looked up and
815 I<ENCODING> is passed to the sub as its first argument. The example
816 is another way to names as used in X11 font names to alias the MIME
817 names for the iso-8859-* family.
821 =head2 Defining Encodings
823 use Encode qw(define_alias);
824 define_encoding( $object, 'canonicalName' [,alias...]);
826 Causes I<canonicalName> to be associated with I<$object>. The object
827 should provide the interface described in L</"IMPLEMENTATION CLASSES">
828 below. If more than two arguments are provided then additional
829 arguments are taken as aliases for I<$object> as for C<define_alias>.
831 =head1 Encoding and IO
833 It is very common to want to do encoding transformations when
834 reading or writing files, network connections, pipes etc.
835 If Perl is configured to use the new 'perlio' IO system then
836 C<Encode> provides a "layer" (See L<perliol>) which can transform
837 data as it is read or written.
839 Here is how the blind poet would modernise the encoding:
842 open(my $iliad,'<:encoding(iso-8859-7)','iliad.greek');
843 open(my $utf8,'>:utf8','iliad.utf8');
849 In addition the new IO system can also be configured to read/write
850 UTF-8 encoded characters (as noted above this is efficient):
852 open(my $fh,'>:utf8','anything');
853 print $fh "Any \x{0021} string \N{SMILEY FACE}\n";
855 Either of the above forms of "layer" specifications can be made the default
856 for a lexical scope with the C<use open ...> pragma. See L<open>.
858 Once a handle is open is layers can be altered using C<binmode>.
860 Without any such configuration, or if Perl itself is built using
861 system's own IO, then write operations assume that file handle accepts
862 only I<bytes> and will C<die> if a character larger than 255 is
863 written to the handle. When reading, each octet from the handle
864 becomes a byte-in-a-character. Note that this default is the same
865 behaviour as bytes-only languages (including Perl before v5.6) would
866 have, and is sufficient to handle native 8-bit encodings
867 e.g. iso-8859-1, EBCDIC etc. and any legacy mechanisms for handling
868 other encodings and binary data.
870 In other cases it is the programs responsibility to transform
871 characters into bytes using the API above before doing writes, and to
872 transform the bytes read from a handle into characters before doing
873 "character operations" (e.g. C<lc>, C</\W+/>, ...).
875 You can also use PerlIO to convert larger amounts of data you don't
876 want to bring into memory. For example to convert between ISO 8859-1
877 (Latin 1) and UTF-8 (or UTF-EBCDIC in EBCDIC machines):
879 open(F, "<:encoding(iso-8859-1)", "data.txt") or die $!;
880 open(G, ">:utf8", "data.utf") or die $!;
881 while (<F>) { print G }
883 # Could also do "print G <F>" but that would pull
884 # the whole file into memory just to write it out again.
888 open(my $f, "<:encoding(cp1252)")
889 open(my $g, ">:encoding(iso-8859-2)")
890 open(my $h, ">:encoding(latin9)") # iso-8859-15
892 See L<PerlIO> for more information.
894 See also L<encoding> for how to change the default encoding of the
897 =head1 Encoding How to ...
903 =item * IO with mixed content (faking iso-2020-*)
905 =item * MIME's Content-Length:
907 =item * UTF-8 strings in binary data.
909 =item * Perl/Encode wrappers on non-Unicode XS modules.
913 =head1 Messing with Perl's Internals
915 The following API uses parts of Perl's internals in the current
916 implementation. As such they are efficient, but may change.
920 =item * is_utf8(STRING [, CHECK])
922 [INTERNAL] Test whether the UTF-8 flag is turned on in the STRING.
923 If CHECK is true, also checks the data in STRING for being well-formed
924 UTF-8. Returns true if successful, false otherwise.
926 =item * valid_utf8(STRING)
928 [INTERNAL] Test whether STRING is in a consistent state. Will return
929 true if string is held as bytes, or is well-formed UTF-8 and has the
930 UTF-8 flag on. Main reason for this routine is to allow Perl's
931 testsuite to check that operations have left strings in a consistent
938 [INTERNAL] Turn on the UTF-8 flag in STRING. The data in STRING is
939 B<not> checked for being well-formed UTF-8. Do not use unless you
940 B<know> that the STRING is well-formed UTF-8. Returns the previous
941 state of the UTF-8 flag (so please don't test the return value as
942 I<not> success or failure), or C<undef> if STRING is not a string.
948 [INTERNAL] Turn off the UTF-8 flag in STRING. Do not use frivolously.
949 Returns the previous state of the UTF-8 flag (so please don't test the
950 return value as I<not> success or failure), or C<undef> if STRING is
955 =head1 IMPLEMENTATION CLASSES
957 As mentioned above encodings are (in the current implementation at least)
958 defined by objects. The mapping of encoding name to object is via the
961 The values of the hash can currently be either strings or objects.
962 The string form may go away in the future. The string form occurs
963 when C<encodings()> has scanned C<@INC> for loadable encodings but has
964 not actually loaded the encoding in question. This is because the
965 current "loading" process is all Perl and a bit slow.
967 Once an encoding is loaded then value of the hash is object which
968 implements the encoding. The object should provide the following
975 Should return the string representing the canonical name of the encoding.
977 =item -E<gt>new_sequence
979 This is a placeholder for encodings with state. It should return an
980 object which implements this interface, all current implementations
981 return the original object.
983 =item -E<gt>encode($string,$check)
985 Should return the octet sequence representing I<$string>. If I<$check>
986 is true it should modify I<$string> in place to remove the converted
987 part (i.e. the whole string unless there is an error). If an error
988 occurs it should return the octet sequence for the fragment of string
989 that has been converted, and modify $string in-place to remove the
990 converted part leaving it starting with the problem fragment.
992 If check is is false then C<encode> should make a "best effort" to
993 convert the string - for example by using a replacement character.
995 =item -E<gt>decode($octets,$check)
997 Should return the string that I<$octets> represents. If I<$check> is
998 true it should modify I<$octets> in place to remove the converted part
999 (i.e. the whole sequence unless there is an error). If an error
1000 occurs it should return the fragment of string that has been
1001 converted, and modify $octets in-place to remove the converted part
1002 leaving it starting with the problem fragment.
1004 If check is is false then C<decode> should make a "best effort" to
1005 convert the string - for example by using Unicode's "\x{FFFD}" as a
1006 replacement character.
1010 It should be noted that the check behaviour is different from the
1011 outer public API. The logic is that the "unchecked" case is useful
1012 when encoding is part of a stream which may be reporting errors
1013 (e.g. STDERR). In such cases it is desirable to get everything
1014 through somehow without causing additional errors which obscure the
1015 original one. Also the encoding is best placed to know what the
1016 correct replacement character is, so if that is the desired behaviour
1017 then letting low level code do it is the most efficient.
1019 In contrast if check is true, the scheme above allows the encoding to
1020 do as much as it can and tell layer above how much that was. What is
1021 lacking at present is a mechanism to report what went wrong. The most
1022 likely interface will be an additional method call to the object, or
1023 perhaps (to avoid forcing per-stream objects on otherwise stateless
1024 encodings) and additional parameter.
1026 It is also highly desirable that encoding classes inherit from
1027 C<Encode::Encoding> as a base class. This allows that class to define
1028 additional behaviour for all encoding objects. For example built in
1029 Unicode, UCS-2 and UTF-8 classes use :
1031 package Encode::MyEncoding;
1032 use base qw(Encode::Encoding);
1034 __PACKAGE__->Define(qw(myCanonical myAlias));
1036 To create an object with bless {Name => ...},$class, and call
1037 define_encoding. They inherit their C<name> method from
1038 C<Encode::Encoding>.
1040 =head2 Compiled Encodings
1042 F<Encode.xs> provides a class C<Encode::XS> which provides the
1043 interface described above. It calls a generic octet-sequence to
1044 octet-sequence "engine" that is driven by tables (defined in
1045 F<encengine.c>). The same engine is used for both encode and
1046 decode. C<Encode:XS>'s C<encode> forces Perl's characters to their
1047 UTF-8 form and then treats them as just another multibyte
1048 encoding. C<Encode:XS>'s C<decode> transforms the sequence and then
1049 turns the UTF-8-ness flag as that is the form that the tables are
1050 defined to produce. For details of the engine see the comments in
1053 The tables are produced by the Perl script F<compile> (the name needs
1054 to change so we can eventually install it somewhere). F<compile> can
1055 currently read two formats:
1061 This is a coined format used by Tcl. It is documented in
1062 Encode/EncodeFormat.pod.
1066 This is the semi-standard format used by IBM's ICU package.
1070 F<compile> can write the following forms:
1076 See above - the F<Encode/*.ucm> files provided with the distribution have
1077 been created from the original Tcl .enc files using this approach.
1081 Produces tables as C data structures - this is used to build in encodings
1082 into F<Encode.so>/F<Encode.dll>.
1086 In theory this allows encodings to be stand-alone loadable Perl
1087 extensions. The process has not yet been tested. The plan is to use
1088 this approach for large East Asian encodings.
1092 The set of encodings built-in to F<Encode.so>/F<Encode.dll> is
1093 determined by F<Makefile.PL>. The current set is as follows:
1097 =item ascii and iso-8859-*
1099 That is all the common 8-bit "western" encodings.
1101 =item IBM-1047 and two other variants of EBCDIC.
1103 These are the same variants that are supported by EBCDIC Perl as
1104 "native" encodings. They are included to prove "reversibility" of
1105 some constructs in EBCDIC Perl.
1107 =item symbol and dingbats as used by Tk on X11.
1109 (The reason Encode got started was to support Perl/Tk.)
1113 That set is rather ad hoc and has been driven by the needs of the
1114 tests rather than the needs of typical applications. It is likely
1119 L<perlunicode>, L<perlebcdic>, L<perlfunc/open>, L<PerlIO>, L<encoding>