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 => '"${1}8"' );
125 # This is a font issue, not an encoding issue.
126 # (The currency symbol of the Latin 1 upper half
127 # has been redefined as the euro symbol.)
128 define_alias( qr/^(.+)\@euro$/i => '"$1"' );
130 # Allow latin-1 style names as well
131 define_alias( qr/^(?:iso[-_]?)?latin[-_]?(\d+)$/i => '"iso-8859-$latin2iso_num[$1]"' );
133 # Allow winlatin1 style names as well
134 define_alias( qr/^win(latin[12]|cyrillic|baltic|greek|turkish|hebrew|arabic|baltic|vietnamese)$/i => '"cp$winlatin2cp{\u$1}"' );
136 # Common names for non-latin prefered MIME names
137 define_alias( 'ascii' => 'US-ascii',
138 'cyrillic' => 'iso-8859-5',
139 'arabic' => 'iso-8859-6',
140 'greek' => 'iso-8859-7',
141 'hebrew' => 'iso-8859-8',
142 'thai' => 'iso-8859-11',
143 'tis620' => 'iso-8859-11',
146 # At least AIX has IBM-NNN (surprisingly...) instead of cpNNN.
147 define_alias( qr/^ibm[-_]?(\d\d\d\d?)$/i => '"cp$1"');
149 # Standardize on the dashed versions.
150 define_alias( qr/^utf8$/i => 'utf-8' );
151 define_alias( qr/^koi8r$/i => 'koi8-r' );
152 define_alias( qr/^koi8u$/i => 'koi8-u' );
154 # TODO: HP-UX '8' encodings arabic8 greek8 hebrew8 kana8 thai8 turkish8
155 # TODO: HP-UX '15' encodings japanese15 korean15 roi15
156 # TODO: Cyrillic encoding ISO-IR-111 (useful?)
157 # TODO: Chinese encodings GB18030 GBK Big5-HSKCS EUC-TW
158 # TODO: Armenian encoding ARMSCII-8
159 # TODO: Hebrew encoding ISO-8859-8-1
160 # TODO: Thai encoding TCVN
161 # TODO: Korean encoding Johab
162 # TODO: Vietnamese encodings VISCII VPS
163 # TODO: Japanese encoding JIS (not the same as SJIS)
164 # TODO: Mac Asian+African encodings: Arabic Armenian Bengali Burmese
165 # ChineseSimp ChineseTrad Devanagari Ethiopic ExtArabic
166 # Farsi Georgian Gujarati Gurmukhi Hebrew Japanese
167 # Kannada Khmer Korean Laotian Malayalam Mongolian
168 # Oriya Sinhalese Symbol Tamil Telugu Tibetan Vietnamese
169 # TODO: what is the Japanese 'UJIS' encoding seen in some Linuxes?
171 # Map white space and _ to '-'
172 define_alias( qr/^(\S+)[\s_]+(.*)$/i => '"$1-$2"' );
178 $encoding{$name} = $obj;
180 define_alias($lc => $obj) unless $lc eq $name;
184 define_alias($alias,$obj);
191 my ($class,$name) = @_;
193 if (ref($name) && $name->can('new_sequence'))
198 if (exists $encoding{$name})
200 return $encoding{$name};
202 if (exists $encoding{$lc})
204 return $encoding{$lc};
207 my $oc = $class->findAlias($name);
208 return $oc if defined $oc;
209 return $class->findAlias($lc) if $lc ne $name;
217 return __PACKAGE__->getEncoding($name);
222 my ($name,$string,$check) = @_;
223 my $enc = find_encoding($name);
224 croak("Unknown encoding '$name'") unless defined $enc;
225 my $octets = $enc->encode($string,$check);
226 return undef if ($check && length($string));
232 my ($name,$octets,$check) = @_;
233 my $enc = find_encoding($name);
234 croak("Unknown encoding '$name'") unless defined $enc;
235 my $string = $enc->decode($octets,$check);
236 $_[1] = $octets if $check;
242 my ($string,$from,$to,$check) = @_;
243 my $f = find_encoding($from);
244 croak("Unknown encoding '$from'") unless defined $f;
245 my $t = find_encoding($to);
246 croak("Unknown encoding '$to'") unless defined $t;
247 my $uni = $f->decode($string,$check);
248 return undef if ($check && length($string));
249 $string = $t->encode($uni,$check);
250 return undef if ($check && length($uni));
251 return length($_[0] = $string);
264 return undef unless utf8::decode($str);
268 package Encode::Encoding;
269 # Base class for classes which implement encodings
274 my $canonical = shift;
275 $obj = bless { Name => $canonical },$obj unless ref $obj;
276 # warn "$canonical => $obj\n";
277 Encode::define_encoding($obj, $canonical, @_);
280 sub name { shift->{'Name'} }
282 # Temporary legacy methods
283 sub toUnicode { shift->decode(@_) }
284 sub fromUnicode { shift->encode(@_) }
286 sub new_sequence { return $_[0] }
289 use base 'Encode::Encoding';
291 package Encode::Internal;
292 use base 'Encode::Encoding';
294 # Dummy package that provides the encode interface but leaves data
295 # as UTF-X encoded. It is here so that from_to() works.
297 __PACKAGE__->Define('Internal');
299 Encode::define_alias( 'Unicode' => 'Internal' ) if ord('A') == 65;
303 my ($obj,$str,$chk) = @_;
311 package Encoding::Unicode;
312 use base 'Encode::Encoding';
314 __PACKAGE__->Define('Unicode') unless ord('A') == 65;
318 my ($obj,$str,$chk) = @_;
320 for (my $i = 0; $i < length($str); $i++)
322 $res .= chr(utf8::unicode_to_native(ord(substr($str,$i,1))));
330 my ($obj,$str,$chk) = @_;
332 for (my $i = 0; $i < length($str); $i++)
334 $res .= chr(utf8::native_to_unicode(ord(substr($str,$i,1))));
341 package Encode::utf8;
342 use base 'Encode::Encoding';
343 # package to allow long-hand
344 # $octets = encode( utf8 => $string );
347 __PACKAGE__->Define(qw(UTF-8 utf8));
351 my ($obj,$octets,$chk) = @_;
352 my $str = Encode::decode_utf8($octets);
363 my ($obj,$string,$chk) = @_;
364 my $octets = Encode::encode_utf8($string);
369 package Encode::iso10646_1;
370 use base 'Encode::Encoding';
371 # Encoding is 16-bit network order Unicode (no surogates)
372 # Used for X font encodings
374 __PACKAGE__->Define(qw(UCS-2 iso-10646-1));
378 my ($obj,$str,$chk) = @_;
382 my $code = unpack('n',substr($str,0,2,'')) & 0xffff;
385 $_[1] = $str if $chk;
392 my ($obj,$uni,$chk) = @_;
396 my $ch = substr($uni,0,1,'');
403 $str .= pack('n',$x);
405 $_[1] = $uni if $chk;
409 package Encode::ucs_2le;
410 use base 'Encode::Encoding';
412 __PACKAGE__->Define(qw(UCS-2le UCS-2LE ucs-2le));
416 my ($obj,$str,$chk) = @_;
420 my $code = unpack('v',substr($str,0,2,'')) & 0xffff;
423 $_[1] = $str if $chk;
430 my ($obj,$uni,$chk) = @_;
434 my $ch = substr($uni,0,1,'');
441 $str .= pack('v',$x);
443 $_[1] = $uni if $chk;
447 # switch back to Encode package in case we ever add AutoLoader
456 Encode - character encodings
464 The C<Encode> module provides the interfaces between Perl's strings
465 and the rest of the system. Perl strings are sequences of B<characters>.
467 The repertoire of characters that Perl can represent is at least that
468 defined by the Unicode Consortium. On most platforms the ordinal
469 values of the characters (as returned by C<ord(ch)>) is the "Unicode
470 codepoint" for the character (the exceptions are those platforms where
471 the legacy encoding is some variant of EBCDIC rather than a super-set
472 of ASCII - see L<perlebcdic>).
474 Traditionaly computer data has been moved around in 8-bit chunks
475 often called "bytes". These chunks are also known as "octets" in
476 networking standards. Perl is widely used to manipulate data of
477 many types - not only strings of characters representing human or
478 computer languages but also "binary" data being the machines representation
479 of numbers, pixels in an image - or just about anything.
481 When Perl is processing "binary data" the programmer wants Perl to process
482 "sequences of bytes". This is not a problem for Perl - as a byte has 256
483 possible values it easily fits in Perl's much larger "logical character".
491 I<character>: a character in the range 0..(2**32-1) (or more).
492 (What Perl's strings are made of.)
496 I<byte>: a character in the range 0..255
497 (A special case of a Perl character.)
501 I<octet>: 8 bits of data, with ordinal values 0..255
502 (Term for bytes passed to or from a non-Perl context, e.g. disk file.)
506 The marker [INTERNAL] marks Internal Implementation Details, in
507 general meant only for those who think they know what they are doing,
508 and such details may change in future releases.
512 =head2 Characteristics of an Encoding
514 An encoding has a "repertoire" of characters that it can represent,
515 and for each representable character there is at least one sequence of
516 octets that represents it.
518 =head2 Types of Encodings
520 Encodings can be divided into the following types:
524 =item * Fixed length 8-bit (or less) encodings.
526 Each character is a single octet so may have a repertoire of up to
527 256 characters. ASCII and iso-8859-* are typical examples.
529 =item * Fixed length 16-bit encodings
531 Each character is two octets so may have a repertoire of up to
532 65 536 characters. Unicode's UCS-2 is an example. Also used for
533 encodings for East Asian languages.
535 =item * Fixed length 32-bit encodings.
537 Not really very "encoded" encodings. The Unicode code points
538 are just represented as 4-octet integers. None the less because
539 different architectures use different representations of integers
540 (so called "endian") there at least two disctinct encodings.
542 =item * Multi-byte encodings
544 The number of octets needed to represent a character varies.
545 UTF-8 is a particularly complex but regular case of a multi-byte
546 encoding. Several East Asian countries use a multi-byte encoding
547 where 1-octet is used to cover western roman characters and Asian
548 characters get 2-octets.
549 (UTF-16 is strictly a multi-byte encoding taking either 2 or 4 octets
550 to represent a Unicode code point.)
552 =item * "Escape" encodings.
554 These encodings embed "escape sequences" into the octet sequence
555 which describe how the following octets are to be interpreted.
556 The iso-2022-* family is typical. Following the escape sequence
557 octets are encoded by an "embedded" encoding (which will be one
558 of the above types) until another escape sequence switches to
559 a different "embedded" encoding.
561 These schemes are very flexible and can handle mixed languages but are
562 very complex to process (and have state). No escape encodings are
563 implemented for Perl yet.
567 =head2 Specifying Encodings
569 Encodings can be specified to the API described below in two ways:
575 Encoding names are strings with characters taken from a restricted
576 repertoire. See L</"Encoding Names">.
578 =item 2. As an object
580 Encoding objects are returned by C<find_encoding($name)>.
584 =head2 Encoding Names
586 Encoding names are case insensitive. White space in names is ignored.
587 In addition an encoding may have aliases. Each encoding has one
588 "canonical" name. The "canonical" name is chosen from the names of
589 the encoding by picking the first in the following sequence:
593 =item * The MIME name as defined in IETF RFC-XXXX.
595 =item * The name in the IANA registry.
597 =item * The name used by the the organization that defined it.
601 Because of all the alias issues, and because in the general case
602 encodings have state C<Encode> uses the encoding object internally
603 once an operation is in progress.
605 =head1 PERL ENCODING API
607 =head2 Generic Encoding Interface
613 $bytes = encode(ENCODING, $string[, CHECK])
615 Encodes string from Perl's internal form into I<ENCODING> and returns
616 a sequence of octets. For CHECK see L</"Handling Malformed Data">.
620 $string = decode(ENCODING, $bytes[, CHECK])
622 Decode sequence of octets assumed to be in I<ENCODING> into Perl's
623 internal form and returns the resulting string. For CHECK see
624 L</"Handling Malformed Data">.
628 from_to($string, FROM_ENCODING, TO_ENCODING[, CHECK])
630 Convert B<in-place> the data between two encodings. How did the data
631 in $string originally get to be in FROM_ENCODING? Either using
632 encode() or through PerlIO: See L</"Encoding and IO">. For CHECK
633 see L</"Handling Malformed Data">.
635 For example to convert ISO 8859-1 data to UTF-8:
637 from_to($data, "iso-8859-1", "utf-8");
639 and to convert it back:
641 from_to($data, "utf-8", "iso-8859-1");
643 Note that because the conversion happens in place, the data to be
644 converted cannot be a string constant, it must be a scalar variable.
648 =head2 Handling Malformed Data
650 If CHECK is not set, C<undef> is returned. If the data is supposed to
651 be UTF-8, an optional lexical warning (category utf8) is given. If
652 CHECK is true but not a code reference, dies.
654 It would desirable to have a way to indicate that transform should use
655 the encodings "replacement character" - no such mechanism is defined yet.
657 It is also planned to allow I<CHECK> to be a code reference.
659 This is not yet implemented as there are design issues with what its
660 arguments should be and how it returns its results.
666 Passed remaining fragment of string being processed.
667 Modifies it in place to remove bytes/characters it can understand
668 and returns a string used to represent them.
672 my $ch = substr($_[0],0,1,'');
673 return sprintf("\x{%02X}",ord($ch);
676 This scheme is close to how underlying C code for Encode works, but gives
677 the fixup routine very little context.
681 Passed original string, and an index into it of the problem area, and
682 output string so far. Appends what it will to output string and
683 returns new index into original string. For example:
686 # my ($s,$i,$d) = @_;
687 my $ch = substr($_[0],$_[1],1);
688 $_[2] .= sprintf("\x{%02X}",ord($ch);
692 This scheme gives maximal control to the fixup routine but is more
693 complicated to code, and may need internals of Encode to be tweaked to
694 keep original string intact.
700 Multiple return values rather than in-place modifications.
702 Index into the string could be pos($str) allowing s/\G...//.
708 The Unicode consortium defines the UTF-8 standard as a way of encoding
709 the entire Unicode repertiore as sequences of octets. This encoding is
710 expected to become very widespread. Perl can use this form internaly
711 to represent strings, so conversions to and from this form are
712 particularly efficient (as octets in memory do not have to change,
713 just the meta-data that tells Perl how to treat them).
719 $bytes = encode_utf8($string);
721 The characters that comprise string are encoded in Perl's superset of UTF-8
722 and the resulting octets returned as a sequence of bytes. All possible
723 characters have a UTF-8 representation so this function cannot fail.
727 $string = decode_utf8($bytes [,CHECK]);
729 The sequence of octets represented by $bytes is decoded from UTF-8
730 into a sequence of logical characters. Not all sequences of octets
731 form valid UTF-8 encodings, so it is possible for this call to fail.
732 For CHECK see L</"Handling Malformed Data">.
736 =head2 Other Encodings of Unicode
738 UTF-16 is similar to UCS-2, 16 bit or 2-byte chunks. UCS-2 can only
739 represent 0..0xFFFF, while UTF-16 has a I<surrogate pair> scheme which
740 allows it to cover the whole Unicode range.
742 Surrogates are code points set aside to encode the 0x01000..0x10FFFF
743 range of Unicode code points in pairs of 16-bit units. The I<high
744 surrogates> are the range 0xD800..0xDBFF, and the I<low surrogates>
745 are the range 0xDC00..0xDFFFF. The surrogate encoding is
747 $hi = ($uni - 0x10000) / 0x400 + 0xD800;
748 $lo = ($uni - 0x10000) % 0x400 + 0xDC00;
752 $uni = 0x10000 + ($hi - 0xD8000) * 0x400 + ($lo - 0xDC00);
754 Encode implements big-endian UCS-2 aliased to "iso-10646-1" as that
755 happens to be the name used by that representation when used with X11
758 UTF-32 or UCS-4 is 32-bit or 4-byte chunks. Perl's logical characters
759 can be considered as being in this form without encoding. An encoding
760 to transfer strings in this form (e.g. to write them to a file) would
763 pack('L*', unpack('U*', $string)); # native
765 pack('V*', unpack('U*', $string)); # little-endian
767 pack('N*', unpack('U*', $string)); # big-endian
769 depending on the endianness required.
771 No UTF-32 encodings are implemented yet.
773 Both UCS-2 and UCS-4 style encodings can have "byte order marks" by
774 representing the code point 0xFFFE as the very first thing in a file.
776 =head2 Listing available encodings
778 use Encode qw(encodings);
781 Returns a list of the canonical names of the available encodings.
783 =head2 Defining Aliases
785 use Encode qw(define_alias);
786 define_alias( newName => ENCODING);
788 Allows newName to be used as am alias for ENCODING. ENCODING may be
789 either the name of an encoding or and encoding object (as above).
791 Currently I<newName> can be specified in the following ways:
795 =item As a simple string.
797 =item As a qr// compiled regular expression, e.g.:
799 define_alias( qr/^iso8859-(\d+)$/i => '"iso-8859-$1"' );
801 In this case if I<ENCODING> is not a reference it is C<eval>-ed to
802 allow C<$1> etc. to be subsituted. The example is one way to names as
803 used in X11 font names to alias the MIME names for the iso-8859-*
806 =item As a code reference, e.g.:
808 define_alias( sub { return /^iso8859-(\d+)$/i ? "iso-8859-$1" : undef } , '');
810 In this case C<$_> will be set to the name that is being looked up and
811 I<ENCODING> is passed to the sub as its first argument. The example
812 is another way to names as used in X11 font names to alias the MIME
813 names for the iso-8859-* family.
817 =head2 Defining Encodings
819 use Encode qw(define_alias);
820 define_encoding( $object, 'canonicalName' [,alias...]);
822 Causes I<canonicalName> to be associated with I<$object>. The object
823 should provide the interface described in L</"IMPLEMENTATION CLASSES">
824 below. If more than two arguments are provided then additional
825 arguments are taken as aliases for I<$object> as for C<define_alias>.
827 =head1 Encoding and IO
829 It is very common to want to do encoding transformations when
830 reading or writing files, network connections, pipes etc.
831 If Perl is configured to use the new 'perlio' IO system then
832 C<Encode> provides a "layer" (See L<perliol>) which can transform
833 data as it is read or written.
835 Here is how the blind poet would modernise the encoding:
838 open(my $iliad,'<:encoding(iso-8859-7)','iliad.greek');
839 open(my $utf8,'>:utf8','iliad.utf8');
845 In addition the new IO system can also be configured to read/write
846 UTF-8 encoded characters (as noted above this is efficient):
848 open(my $fh,'>:utf8','anything');
849 print $fh "Any \x{0021} string \N{SMILEY FACE}\n";
851 Either of the above forms of "layer" specifications can be made the default
852 for a lexical scope with the C<use open ...> pragma. See L<open>.
854 Once a handle is open is layers can be altered using C<binmode>.
856 Without any such configuration, or if Perl itself is built using
857 system's own IO, then write operations assume that file handle accepts
858 only I<bytes> and will C<die> if a character larger than 255 is
859 written to the handle. When reading, each octet from the handle
860 becomes a byte-in-a-character. Note that this default is the same
861 behaviour as bytes-only languages (including Perl before v5.6) would
862 have, and is sufficient to handle native 8-bit encodings
863 e.g. iso-8859-1, EBCDIC etc. and any legacy mechanisms for handling
864 other encodings and binary data.
866 In other cases it is the programs responsibility to transform
867 characters into bytes using the API above before doing writes, and to
868 transform the bytes read from a handle into characters before doing
869 "character operations" (e.g. C<lc>, C</\W+/>, ...).
871 You can also use PerlIO to convert larger amounts of data you don't
872 want to bring into memory. For example to convert between ISO 8859-1
873 (Latin 1) and UTF-8 (or UTF-EBCDIC in EBCDIC machines):
875 open(F, "<:encoding(iso-8859-1)", "data.txt") or die $!;
876 open(G, ">:utf8", "data.utf") or die $!;
877 while (<F>) { print G }
879 # Could also do "print G <F>" but that would pull
880 # the whole file into memory just to write it out again.
884 open(my $f, "<:encoding(cp1252)")
885 open(my $g, ">:encoding(iso-8859-2)")
886 open(my $h, ">:encoding(latin9)") # iso-8859-15
888 See L<PerlIO> for more information.
890 See also L<encoding> for how to change the default encoding of the
893 =head1 Encoding How to ...
899 =item * IO with mixed content (faking iso-2020-*)
901 =item * MIME's Content-Length:
903 =item * UTF-8 strings in binary data.
905 =item * Perl/Encode wrappers on non-Unicode XS modules.
909 =head1 Messing with Perl's Internals
911 The following API uses parts of Perl's internals in the current
912 implementation. As such they are efficient, but may change.
916 =item * is_utf8(STRING [, CHECK])
918 [INTERNAL] Test whether the UTF-8 flag is turned on in the STRING.
919 If CHECK is true, also checks the data in STRING for being well-formed
920 UTF-8. Returns true if successful, false otherwise.
922 =item * valid_utf8(STRING)
924 [INTERNAL] Test whether STRING is in a consistent state. Will return
925 true if string is held as bytes, or is well-formed UTF-8 and has the
926 UTF-8 flag on. Main reason for this routine is to allow Perl's
927 testsuite to check that operations have left strings in a consistent
934 [INTERNAL] Turn on the UTF-8 flag in STRING. The data in STRING is
935 B<not> checked for being well-formed UTF-8. Do not use unless you
936 B<know> that the STRING is well-formed UTF-8. Returns the previous
937 state of the UTF-8 flag (so please don't test the return value as
938 I<not> success or failure), or C<undef> if STRING is not a string.
944 [INTERNAL] Turn off the UTF-8 flag in STRING. Do not use frivolously.
945 Returns the previous state of the UTF-8 flag (so please don't test the
946 return value as I<not> success or failure), or C<undef> if STRING is
951 =head1 IMPLEMENTATION CLASSES
953 As mentioned above encodings are (in the current implementation at least)
954 defined by objects. The mapping of encoding name to object is via the
957 The values of the hash can currently be either strings or objects.
958 The string form may go away in the future. The string form occurs
959 when C<encodings()> has scanned C<@INC> for loadable encodings but has
960 not actually loaded the encoding in question. This is because the
961 current "loading" process is all Perl and a bit slow.
963 Once an encoding is loaded then value of the hash is object which
964 implements the encoding. The object should provide the following
971 Should return the string representing the canonical name of the encoding.
973 =item -E<gt>new_sequence
975 This is a placeholder for encodings with state. It should return an
976 object which implements this interface, all current implementations
977 return the original object.
979 =item -E<gt>encode($string,$check)
981 Should return the octet sequence representing I<$string>. If I<$check>
982 is true it should modify I<$string> in place to remove the converted
983 part (i.e. the whole string unless there is an error). If an error
984 occurs it should return the octet sequence for the fragment of string
985 that has been converted, and modify $string in-place to remove the
986 converted part leaving it starting with the problem fragment.
988 If check is is false then C<encode> should make a "best effort" to
989 convert the string - for example by using a replacement character.
991 =item -E<gt>decode($octets,$check)
993 Should return the string that I<$octets> represents. If I<$check> is
994 true it should modify I<$octets> in place to remove the converted part
995 (i.e. the whole sequence unless there is an error). If an error
996 occurs it should return the fragment of string that has been
997 converted, and modify $octets in-place to remove the converted part
998 leaving it starting with the problem fragment.
1000 If check is is false then C<decode> should make a "best effort" to
1001 convert the string - for example by using Unicode's "\x{FFFD}" as a
1002 replacement character.
1006 It should be noted that the check behaviour is different from the
1007 outer public API. The logic is that the "unchecked" case is useful
1008 when encoding is part of a stream which may be reporting errors
1009 (e.g. STDERR). In such cases it is desirable to get everything
1010 through somehow without causing additional errors which obscure the
1011 original one. Also the encoding is best placed to know what the
1012 correct replacement character is, so if that is the desired behaviour
1013 then letting low level code do it is the most efficient.
1015 In contrast if check is true, the scheme above allows the encoding to
1016 do as much as it can and tell layer above how much that was. What is
1017 lacking at present is a mechanism to report what went wrong. The most
1018 likely interface will be an additional method call to the object, or
1019 perhaps (to avoid forcing per-stream objects on otherwise stateless
1020 encodings) and additional parameter.
1022 It is also highly desirable that encoding classes inherit from
1023 C<Encode::Encoding> as a base class. This allows that class to define
1024 additional behaviour for all encoding objects. For example built in
1025 Unicode, UCS-2 and UTF-8 classes use :
1027 package Encode::MyEncoding;
1028 use base qw(Encode::Encoding);
1030 __PACKAGE__->Define(qw(myCanonical myAlias));
1032 To create an object with bless {Name => ...},$class, and call
1033 define_encoding. They inherit their C<name> method from
1034 C<Encode::Encoding>.
1036 =head2 Compiled Encodings
1038 F<Encode.xs> provides a class C<Encode::XS> which provides the
1039 interface described above. It calls a generic octet-sequence to
1040 octet-sequence "engine" that is driven by tables (defined in
1041 F<encengine.c>). The same engine is used for both encode and
1042 decode. C<Encode:XS>'s C<encode> forces Perl's characters to their
1043 UTF-8 form and then treats them as just another multibyte
1044 encoding. C<Encode:XS>'s C<decode> transforms the sequence and then
1045 turns the UTF-8-ness flag as that is the form that the tables are
1046 defined to produce. For details of the engine see the comments in
1049 The tables are produced by the Perl script F<compile> (the name needs
1050 to change so we can eventually install it somewhere). F<compile> can
1051 currently read two formats:
1057 This is a coined format used by Tcl. It is documented in
1058 Encode/EncodeFormat.pod.
1062 This is the semi-standard format used by IBM's ICU package.
1066 F<compile> can write the following forms:
1072 See above - the F<Encode/*.ucm> files provided with the distribution have
1073 been created from the original Tcl .enc files using this approach.
1077 Produces tables as C data structures - this is used to build in encodings
1078 into F<Encode.so>/F<Encode.dll>.
1082 In theory this allows encodings to be stand-alone loadable Perl
1083 extensions. The process has not yet been tested. The plan is to use
1084 this approach for large East Asian encodings.
1088 The set of encodings built-in to F<Encode.so>/F<Encode.dll> is
1089 determined by F<Makefile.PL>. The current set is as follows:
1093 =item ascii and iso-8859-*
1095 That is all the common 8-bit "western" encodings.
1097 =item IBM-1047 and two other variants of EBCDIC.
1099 These are the same variants that are supported by EBCDIC Perl as
1100 "native" encodings. They are included to prove "reversibility" of
1101 some constructs in EBCDIC Perl.
1103 =item symbol and dingbats as used by Tk on X11.
1105 (The reason Encode got started was to support Perl/Tk.)
1109 That set is rather ad hoc and has been driven by the needs of the
1110 tests rather than the needs of typical applications. It is likely
1115 L<perlunicode>, L<perlebcdic>, L<perlfunc/open>, L<PerlIO>, L<encoding>