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
49 return keys %encoding;
56 unless (exists $alias{$_})
58 for (my $i=0; $i < @alias; $i += 2)
60 my $alias = $alias[$i];
61 my $val = $alias[$i+1];
64 if (ref($alias) eq 'Regexp' && $_ =~ $alias)
68 elsif (ref($alias) eq 'CODE')
70 $new = &{$alias}($val)
72 elsif (lc($_) eq lc($alias))
78 next if $new eq $_; # avoid (direct) recursion on bugs
79 my $enc = (ref($new)) ? $new : find_encoding($new);
95 my ($alias,$name) = splice(@_,0,2);
96 push(@alias, $alias => $name);
100 # Allow variants of iso-8859-1 etc.
101 define_alias( qr/^iso[-_]?(\d+)[-_](\d+)$/i => '"iso-$1-$2"' );
103 # Allow latin-1 style names as well
104 # 0 1 2 3 4 5 6 7 8 9 10
105 my @latin2iso_num = ( 0, 1, 2, 3, 4, 9, 10, 13, 14, 15, 16 );
106 define_alias( qr/^latin[-_]?(\d+)$/i => '"iso-8859-$latin2iso_num[$1]"' );
108 # Common names for non-latin prefered MIME names
109 define_alias( 'ascii' => 'US-ascii',
110 'cyrillic' => 'iso-8859-5',
111 'arabic' => 'iso-8859-6',
112 'greek' => 'iso-8859-7',
113 'hebrew' => 'iso-8859-8');
115 define_alias( 'ibm-1047' => 'cp1047');
117 # Map white space and _ to '-'
118 define_alias( qr/^(\S+)[\s_]+(.*)$/i => '"$1-$2"' );
124 $encoding{$name} = $obj;
126 define_alias($lc => $obj) unless $lc eq $name;
130 define_alias($alias,$obj);
137 my ($class,$name) = @_;
139 if (exists $encoding{$name})
141 return $encoding{$name};
145 return $class->findAlias($name);
152 return __PACKAGE__->getEncoding($name);
157 my ($name,$string,$check) = @_;
158 my $enc = find_encoding($name);
159 croak("Unknown encoding '$name'") unless defined $enc;
160 my $octets = $enc->encode($string,$check);
161 return undef if ($check && length($string));
167 my ($name,$octets,$check) = @_;
168 my $enc = find_encoding($name);
169 croak("Unknown encoding '$name'") unless defined $enc;
170 my $string = $enc->decode($octets,$check);
171 return undef if ($check && length($octets));
177 my ($string,$from,$to,$check) = @_;
178 my $f = find_encoding($from);
179 croak("Unknown encoding '$from'") unless defined $f;
180 my $t = find_encoding($to);
181 croak("Unknown encoding '$to'") unless defined $t;
182 my $uni = $f->decode($string,$check);
183 return undef if ($check && length($string));
184 $string = $t->encode($uni,$check);
185 return undef if ($check && length($uni));
186 return length($_[0] = $string);
199 return undef unless utf8::decode($str);
203 package Encode::Encoding;
204 # Base class for classes which implement encodings
209 my $canonical = shift;
210 $obj = bless { Name => $canonical },$obj unless ref $obj;
211 # warn "$canonical => $obj\n";
212 Encode::define_encoding($obj, $canonical, @_);
215 sub name { shift->{'Name'} }
217 # Temporary legacy methods
218 sub toUnicode { shift->decode(@_) }
219 sub fromUnicode { shift->encode(@_) }
221 sub new_sequence { return $_[0] }
224 use base 'Encode::Encoding';
226 package Encode::Internal;
227 use base 'Encode::Encoding';
229 # Dummy package that provides the encode interface but leaves data
230 # as UTF-X encoded. It is here so that from_to() works.
232 __PACKAGE__->Define('Internal');
234 Encode::define_alias( 'Unicode' => 'Internal' ) if ord('A') == 65;
238 my ($obj,$str,$chk) = @_;
246 package Encoding::Unicode;
247 use base 'Encode::Encoding';
249 __PACKAGE__->Define('Unicode') unless ord('A') == 65;
253 my ($obj,$str,$chk) = @_;
255 for (my $i = 0; $i < length($str); $i++)
257 $res .= chr(utf8::unicode_to_native(ord(substr($str,$i,1))));
265 my ($obj,$str,$chk) = @_;
267 for (my $i = 0; $i < length($str); $i++)
269 $res .= chr(utf8::native_to_unicode(ord(substr($str,$i,1))));
276 package Encode::utf8;
277 use base 'Encode::Encoding';
278 # package to allow long-hand
279 # $octets = encode( utf8 => $string );
282 __PACKAGE__->Define(qw(UTF-8 utf8));
286 my ($obj,$octets,$chk) = @_;
287 my $str = Encode::decode_utf8($octets);
298 my ($obj,$string,$chk) = @_;
299 my $octets = Encode::encode_utf8($string);
304 package Encode::iso10646_1;
305 use base 'Encode::Encoding';
306 # Encoding is 16-bit network order Unicode (no surogates)
307 # Used for X font encodings
309 __PACKAGE__->Define(qw(UCS-2 iso-10646-1));
313 my ($obj,$str,$chk) = @_;
317 my $code = unpack('n',substr($str,0,2,'')) & 0xffff;
320 $_[1] = $str if $chk;
327 my ($obj,$uni,$chk) = @_;
331 my $ch = substr($uni,0,1,'');
338 $str .= pack('n',$x);
340 $_[1] = $uni if $chk;
344 # switch back to Encode package in case we ever add AutoLoader
353 Encode - character encodings
361 The C<Encode> module provides the interfaces between perl's strings
362 and the rest of the system. Perl strings are sequences of B<characters>.
364 The repertoire of characters that Perl can represent is at least that
365 defined by the Unicode Consortium. On most platforms the ordinal values
366 of the characters (as returned by C<ord(ch)>) is the "Unicode codepoint" for
367 the character (the exceptions are those platforms where the legacy
368 encoding is some variant of EBCDIC rather than a super-set of ASCII
369 - see L<perlebcdic>).
371 Traditionaly computer data has been moved around in 8-bit chunks
372 often called "bytes". These chunks are also known as "octets" in
373 networking standards. Perl is widely used to manipulate data of
374 many types - not only strings of characters representing human or
375 computer languages but also "binary" data being the machines representation
376 of numbers, pixels in an image - or just about anything.
378 When perl is processing "binary data" the programmer wants perl to process
379 "sequences of bytes". This is not a problem for perl - as a byte has 256
380 possible values it easily fits in perl's much larger "logical character".
388 I<character>: a character in the range 0..(2**32-1) (or more).
389 (What perl's strings are made of.)
393 I<byte>: a character in the range 0..255
394 (A special case of a perl character.)
398 I<octet>: 8 bits of data, with ordinal values 0..255
399 (Term for bytes passed to or from a non-perl context, e.g. disk file.)
403 The marker [INTERNAL] marks Internal Implementation Details, in
404 general meant only for those who think they know what they are doing,
405 and such details may change in future releases.
409 =head2 Characteristics of an Encoding
411 An encoding has a "repertoire" of characters that it can represent,
412 and for each representable character there is at least one sequence of
413 octets that represents it.
415 =head2 Types of Encodings
417 Encodings can be divided into the following types:
421 =item * Fixed length 8-bit (or less) encodings.
423 Each character is a single octet so may have a repertoire of up to
424 256 characters. ASCII and iso-8859-* are typical examples.
426 =item * Fixed length 16-bit encodings
428 Each character is two octets so may have a repertoire of up to
429 65,536 characters. Unicode's UCS-2 is an example. Also used for
430 encodings for East Asian languages.
432 =item * Fixed length 32-bit encodings.
434 Not really very "encoded" encodings. The Unicode code points
435 are just represented as 4-octet integers. None the less because
436 different architectures use different representations of integers
437 (so called "endian") there at least two disctinct encodings.
439 =item * Multi-byte encodings
441 The number of octets needed to represent a character varies.
442 UTF-8 is a particularly complex but regular case of a multi-byte
443 encoding. Several East Asian countries use a multi-byte encoding
444 where 1-octet is used to cover western roman characters and Asian
445 characters get 2-octets.
446 (UTF-16 is strictly a multi-byte encoding taking either 2 or 4 octets
447 to represent a Unicode code point.)
449 =item * "Escape" encodings.
451 These encodings embed "escape sequences" into the octet sequence
452 which describe how the following octets are to be interpreted.
453 The iso-2022-* family is typical. Following the escape sequence
454 octets are encoded by an "embedded" encoding (which will be one
455 of the above types) until another escape sequence switches to
456 a different "embedded" encoding.
458 These schemes are very flexible and can handle mixed languages but are
459 very complex to process (and have state).
460 No escape encodings are implemented for perl yet.
464 =head2 Specifying Encodings
466 Encodings can be specified to the API described below in two ways:
472 Encoding names are strings with characters taken from a restricted repertoire.
473 See L</"Encoding Names">.
475 =item 2. As an object
477 Encoding objects are returned by C<find_encoding($name)>.
481 =head2 Encoding Names
483 Encoding names are case insensitive. White space in names is ignored.
484 In addition an encoding may have aliases. Each encoding has one "canonical" name.
485 The "canonical" name is chosen from the names of the encoding by picking
486 the first in the following sequence:
490 =item * The MIME name as defined in IETF RFC-XXXX.
492 =item * The name in the IANA registry.
494 =item * The name used by the the organization that defined it.
498 Because of all the alias issues, and because in the general case
499 encodings have state C<Encode> uses the encoding object internally
500 once an operation is in progress.
502 =head1 PERL ENCODING API
504 =head2 Generic Encoding Interface
510 $bytes = encode(ENCODING, $string[, CHECK])
512 Encodes string from perl's internal form into I<ENCODING> and returns a
514 See L</"Handling Malformed Data">.
518 $string = decode(ENCODING, $bytes[, CHECK])
520 Decode sequence of octets assumed to be in I<ENCODING> into perls internal
521 form and returns the resuting string.
522 See L</"Handling Malformed Data">.
526 =head2 Handling Malformed Data
528 If CHECK is not set, C<undef> is returned. If the data is supposed to
529 be UTF-8, an optional lexical warning (category utf8) is given.
530 If CHECK is true but not a code reference, dies.
532 It would desirable to have a way to indicate that transform should use the
533 encodings "replacement character" - no such mechanism is defined yet.
535 It is also planned to allow I<CHECK> to be a code reference.
537 This is not yet implemented as there are design issues with what its arguments
538 should be and how it returns its results.
544 Passed remaining fragment of string being processed.
545 Modifies it in place to remove bytes/characters it can understand
546 and returns a string used to represent them.
550 my $ch = substr($_[0],0,1,'');
551 return sprintf("\x{%02X}",ord($ch);
554 This scheme is close to how underlying C code for Encode works, but gives
555 the fixup routine very little context.
559 Passed original string, and an index into it of the problem area,
560 and output string so far.
561 Appends what it will to output string and returns new index into
566 # my ($s,$i,$d) = @_;
567 my $ch = substr($_[0],$_[1],1);
568 $_[2] .= sprintf("\x{%02X}",ord($ch);
572 This scheme gives maximal control to the fixup routine but is more complicated
573 to code, and may need internals of Encode to be tweaked to keep original
580 Multiple return values rather than in-place modifications.
582 Index into the string could be pos($str) allowing s/\G...//.
588 The Unicode consortium defines the UTF-8 standard as a way of encoding
589 the entire Unicode repertiore as sequences of octets. This encoding
590 is expected to become very widespread. Perl can use this form internaly
591 to represent strings, so conversions to and from this form are particularly
592 efficient (as octets in memory do not have to change, just the meta-data
593 that tells perl how to treat them).
599 $bytes = encode_utf8($string);
601 The characters that comprise string are encoded in perl's superset of UTF-8
602 and the resulting octets returned as a sequence of bytes. All possible
603 characters have a UTF-8 representation so this function cannot fail.
607 $string = decode_utf8($bytes [,CHECK]);
609 The sequence of octets represented by $bytes is decoded from UTF-8 into
610 a sequence of logical characters. Not all sequences of octets form valid
611 UTF-8 encodings, so it is possible for this call to fail.
612 See L</"Handling Malformed Data">.
616 =head2 Other Encodings of Unicode
618 UTF-16 is similar to UCS-2, 16 bit or 2-byte chunks.
619 UCS-2 can only represent 0..0xFFFF, while UTF-16 has a "surogate pair"
620 scheme which allows it to cover the whole Unicode range.
622 Encode implements big-endian UCS-2 aliased to "iso-10646-1" as that
623 happens to be the name used by that representation when used with X11 fonts.
625 UTF-32 or UCS-4 is 32-bit or 4-byte chunks. Perl's logical characters
626 can be considered as being in this form without encoding. An encoding
627 to transfer strings in this form (e.g. to write them to a file) would need to
629 pack('L',map(chr($_),split(//,$string))); # native
631 pack('V',map(chr($_),split(//,$string))); # little-endian
633 pack('N',map(chr($_),split(//,$string))); # big-endian
635 depending on the endian required.
637 No UTF-32 encodings are implemented yet.
639 Both UCS-2 and UCS-4 style encodings can have "byte order marks" by representing
640 the code point 0xFFFE as the very first thing in a file.
642 =head2 Listing available encodings
644 use Encode qw(encodings);
647 Returns a list of the canonical names of the available encodings.
649 =head2 Defining Aliases
651 use Encode qw(define_alias);
652 define_alias( newName => ENCODING);
654 Allows newName to be used as am alias for ENCODING. ENCODING may be either the
655 name of an encoding or and encoding object (as above).
657 Currently I<newName> can be specified in the following ways:
661 =item As a simple string.
663 =item As a qr// compiled regular expression, e.g.:
665 define_alias( qr/^iso8859-(\d+)$/i => '"iso-8859-$1"' );
667 In this case if I<ENCODING> is not a reference it is C<eval>-ed to allow
668 C<$1> etc. to be subsituted.
669 The example is one way to names as used in X11 font names to alias the MIME names for the
672 =item As a code reference, e.g.:
674 define_alias( sub { return /^iso8859-(\d+)$/i ? "iso-8859-$1" : undef } , '');
676 In this case C<$_> will be set to the name that is being looked up and
677 I<ENCODING> is passed to the sub as its first argument.
678 The example is another way to names as used in X11 font names to alias the MIME names for
679 the iso-8859-* family.
683 =head2 Defining Encodings
685 use Encode qw(define_alias);
686 define_encoding( $object, 'canonicalName' [,alias...]);
688 Causes I<canonicalName> to be associated with I<$object>.
689 The object should provide the interface described in L</"IMPLEMENTATION CLASSES"> below.
690 If more than two arguments are provided then additional arguments are taken
691 as aliases for I<$object> as for C<define_alias>.
693 =head1 Encoding and IO
695 It is very common to want to do encoding transformations when
696 reading or writing files, network connections, pipes etc.
697 If perl is configured to use the new 'perlio' IO system then
698 C<Encode> provides a "layer" (See L<perliol>) which can transform
699 data as it is read or written.
701 open(my $ilyad,'>:encoding(iso-8859-7)','ilyad.greek');
704 In addition the new IO system can also be configured to read/write
705 UTF-8 encoded characters (as noted above this is efficient):
707 open(my $fh,'>:utf8','anything');
708 print $fh "Any \x{0021} string \N{SMILEY FACE}\n";
710 Either of the above forms of "layer" specifications can be made the default
711 for a lexical scope with the C<use open ...> pragma. See L<open>.
713 Once a handle is open is layers can be altered using C<binmode>.
715 Without any such configuration, or if perl itself is built using
716 system's own IO, then write operations assume that file handle accepts
717 only I<bytes> and will C<die> if a character larger than 255 is
718 written to the handle. When reading, each octet from the handle
719 becomes a byte-in-a-character. Note that this default is the same
720 behaviour as bytes-only languages (including perl before v5.6) would have,
721 and is sufficient to handle native 8-bit encodings e.g. iso-8859-1,
722 EBCDIC etc. and any legacy mechanisms for handling other encodings
725 In other cases it is the programs responsibility
726 to transform characters into bytes using the API above before
727 doing writes, and to transform the bytes read from a handle into characters
728 before doing "character operations" (e.g. C<lc>, C</\W+/>, ...).
730 =head1 Encoding How to ...
736 =item * IO with mixed content (faking iso-2020-*)
738 =item * MIME's Content-Length:
740 =item * UTF-8 strings in binary data.
742 =item * perl/Encode wrappers on non-Unicode XS modules.
746 =head1 Messing with Perl's Internals
748 The following API uses parts of perl's internals in the current implementation.
749 As such they are efficient, but may change.
753 =item * is_utf8(STRING [, CHECK])
755 [INTERNAL] Test whether the UTF-8 flag is turned on in the STRING.
756 If CHECK is true, also checks the data in STRING for being
757 well-formed UTF-8. Returns true if successful, false otherwise.
759 =item * valid_utf8(STRING)
761 [INTERNAL] Test whether STRING is in a consistent state.
762 Will return true if string is held as bytes, or is well-formed UTF-8
763 and has the UTF-8 flag on.
764 Main reason for this routine is to allow perl's testsuite to check
765 that operations have left strings in a consistent state.
771 [INTERNAL] Turn on the UTF-8 flag in STRING. The data in STRING is
772 B<not> checked for being well-formed UTF-8. Do not use unless you
773 B<know> that the STRING is well-formed UTF-8. Returns the previous
774 state of the UTF-8 flag (so please don't test the return value as
775 I<not> success or failure), or C<undef> if STRING is not a string.
781 [INTERNAL] Turn off the UTF-8 flag in STRING. Do not use frivolously.
782 Returns the previous state of the UTF-8 flag (so please don't test the
783 return value as I<not> success or failure), or C<undef> if STRING is
788 =head1 IMPLEMENTATION CLASSES
790 As mentioned above encodings are (in the current implementation at least)
791 defined by objects. The mapping of encoding name to object is via the
794 The values of the hash can currently be either strings or objects.
795 The string form may go away in the future. The string form occurs
796 when C<encodings()> has scanned C<@INC> for loadable encodings but has
797 not actually loaded the encoding in question. This is because the
798 current "loading" process is all perl and a bit slow.
800 Once an encoding is loaded then value of the hash is object which implements
801 the encoding. The object should provide the following interface:
807 Should return the string representing the canonical name of the encoding.
809 =item -E<gt>new_sequence
811 This is a placeholder for encodings with state. It should return an object
812 which implements this interface, all current implementations return the
815 =item -E<gt>encode($string,$check)
817 Should return the octet sequence representing I<$string>. If I<$check> is true
818 it should modify I<$string> in place to remove the converted part (i.e.
819 the whole string unless there is an error).
820 If an error occurs it should return the octet sequence for the
821 fragment of string that has been converted, and modify $string in-place
822 to remove the converted part leaving it starting with the problem fragment.
824 If check is is false then C<encode> should make a "best effort" to convert
825 the string - for example by using a replacement character.
827 =item -E<gt>decode($octets,$check)
829 Should return the string that I<$octets> represents. If I<$check> is true
830 it should modify I<$octets> in place to remove the converted part (i.e.
831 the whole sequence unless there is an error).
832 If an error occurs it should return the fragment of string
833 that has been converted, and modify $octets in-place to remove the converted part
834 leaving it starting with the problem fragment.
836 If check is is false then C<decode> should make a "best effort" to convert
837 the string - for example by using Unicode's "\x{FFFD}" as a replacement character.
841 It should be noted that the check behaviour is different from the outer
842 public API. The logic is that the "unchecked" case is useful when
843 encoding is part of a stream which may be reporting errors (e.g. STDERR).
844 In such cases it is desirable to get everything through somehow without
845 causing additional errors which obscure the original one. Also the encoding
846 is best placed to know what the correct replacement character is, so if that
847 is the desired behaviour then letting low level code do it is the most efficient.
849 In contrast if check is true, the scheme above allows the encoding to do as
850 much as it can and tell layer above how much that was. What is lacking
851 at present is a mechanism to report what went wrong. The most likely interface
852 will be an additional method call to the object, or perhaps
853 (to avoid forcing per-stream objects on otherwise stateless encodings)
854 and additional parameter.
856 It is also highly desirable that encoding classes inherit from C<Encode::Encoding>
857 as a base class. This allows that class to define additional behaviour for
858 all encoding objects. For example built in Unicode, UCS-2 and UTF-8 classes
861 package Encode::MyEncoding;
862 use base qw(Encode::Encoding);
864 __PACKAGE__->Define(qw(myCanonical myAlias));
866 To create an object with bless {Name => ...},$class, and call define_encoding.
867 They inherit their C<name> method from C<Encode::Encoding>.
869 =head2 Compiled Encodings
871 F<Encode.xs> provides a class C<Encode::XS> which provides the interface described
872 above. It calls a generic octet-sequence to octet-sequence "engine" that is
873 driven by tables (defined in F<encengine.c>). The same engine is used for both
874 encode and decode. C<Encode:XS>'s C<encode> forces perl's characters to their UTF-8 form
875 and then treats them as just another multibyte encoding. C<Encode:XS>'s C<decode> transforms
876 the sequence and then turns the UTF-8-ness flag as that is the form that the tables
877 are defined to produce. For details of the engine see the comments in F<encengine.c>.
879 The tables are produced by the perl script F<compile> (the name needs to change so
880 we can eventually install it somewhere). F<compile> can currently read two formats:
886 This is a coined format used by Tcl. It is documented in Encode/EncodeFormat.pod.
890 This is the semi-standard format used by IBM's ICU package.
894 F<compile> can write the following forms:
900 See above - the F<Encode/*.ucm> files provided with the distribution have
901 been created from the original Tcl .enc files using this approach.
905 Produces tables as C data structures - this is used to build in encodings
906 into F<Encode.so>/F<Encode.dll>.
910 In theory this allows encodings to be stand-alone loadable perl extensions.
911 The process has not yet been tested. The plan is to use this approach
912 for large East Asian encodings.
916 The set of encodings built-in to F<Encode.so>/F<Encode.dll> is determined by
917 F<Makefile.PL>. The current set is as follows:
921 =item ascii and iso-8859-*
923 That is all the common 8-bit "western" encodings.
925 =item IBM-1047 and two other variants of EBCDIC.
927 These are the same variants that are supported by EBCDIC perl as "native" encodings.
928 They are included to prove "reversibility" of some constructs in EBCDIC perl.
930 =item symbol and dingbats as used by Tk on X11.
932 (The reason Encode got started was to support perl/Tk.)
936 That set is rather ad. hoc. and has been driven by the needs of the tests rather
937 than the needs of typical applications. It is likely to be rationalized.
941 L<perlunicode>, L<perlebcdic>, L<perlfunc/open>