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
366 values of the characters (as returned by C<ord(ch)>) is the "Unicode
367 codepoint" for the character (the exceptions are those platforms where
368 the legacy encoding is some variant of EBCDIC rather than a super-set
369 of ASCII - 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). No escape encodings are
460 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
473 repertoire. 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
485 "canonical" name. The "canonical" name is chosen from the names of
486 the encoding by picking 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
513 a sequence of octets. For CHECK see L</"Handling Malformed Data">.
517 $string = decode(ENCODING, $bytes[, CHECK])
519 Decode sequence of octets assumed to be in I<ENCODING> into Perl's
520 internal form and returns the resulting string. For CHECK see
521 L</"Handling Malformed Data">.
525 from_to($string, FROM_ENCODING, TO_ENCODING[, CHECK])
527 Convert B<in-place> the data between two encodings. How did the data
528 in $string originally get to be in FROM_ENCODING? Either using
529 encode() or through PerlIO: See L</"Encode and PerlIO">. For CHECK
530 see L</"Handling Malformed Data">.
532 For example to convert ISO 8859-1 data to UTF-8:
534 from_to($data, "iso-8859-1", "utf-8");
536 and to convert it back:
538 from_to($data, "utf-8", "iso-8859-1");
542 =head2 Handling Malformed Data
544 If CHECK is not set, C<undef> is returned. If the data is supposed to
545 be UTF-8, an optional lexical warning (category utf8) is given. If
546 CHECK is true but not a code reference, dies.
548 It would desirable to have a way to indicate that transform should use
549 the encodings "replacement character" - no such mechanism is defined yet.
551 It is also planned to allow I<CHECK> to be a code reference.
553 This is not yet implemented as there are design issues with what its
554 arguments should be and how it returns its results.
560 Passed remaining fragment of string being processed.
561 Modifies it in place to remove bytes/characters it can understand
562 and returns a string used to represent them.
566 my $ch = substr($_[0],0,1,'');
567 return sprintf("\x{%02X}",ord($ch);
570 This scheme is close to how underlying C code for Encode works, but gives
571 the fixup routine very little context.
575 Passed original string, and an index into it of the problem area, and
576 output string so far. Appends what it will to output string and
577 returns new index into original string. For example:
580 # my ($s,$i,$d) = @_;
581 my $ch = substr($_[0],$_[1],1);
582 $_[2] .= sprintf("\x{%02X}",ord($ch);
586 This scheme gives maximal control to the fixup routine but is more
587 complicated to code, and may need internals of Encode to be tweaked to
588 keep original string intact.
594 Multiple return values rather than in-place modifications.
596 Index into the string could be pos($str) allowing s/\G...//.
602 The Unicode consortium defines the UTF-8 standard as a way of encoding
603 the entire Unicode repertiore as sequences of octets. This encoding is
604 expected to become very widespread. Perl can use this form internaly
605 to represent strings, so conversions to and from this form are
606 particularly efficient (as octets in memory do not have to change,
607 just the meta-data that tells Perl how to treat them).
613 $bytes = encode_utf8($string);
615 The characters that comprise string are encoded in Perl's superset of UTF-8
616 and the resulting octets returned as a sequence of bytes. All possible
617 characters have a UTF-8 representation so this function cannot fail.
621 $string = decode_utf8($bytes [,CHECK]);
623 The sequence of octets represented by $bytes is decoded from UTF-8
624 into a sequence of logical characters. Not all sequences of octets
625 form valid UTF-8 encodings, so it is possible for this call to fail.
626 For CHECK see L</"Handling Malformed Data">.
630 =head2 Other Encodings of Unicode
632 UTF-16 is similar to UCS-2, 16 bit or 2-byte chunks. UCS-2 can only
633 represent 0..0xFFFF, while UTF-16 has a "surrogate pair" scheme which
634 allows it to cover the whole Unicode range.
636 Encode implements big-endian UCS-2 aliased to "iso-10646-1" as that
637 happens to be the name used by that representation when used with X11
640 UTF-32 or UCS-4 is 32-bit or 4-byte chunks. Perl's logical characters
641 can be considered as being in this form without encoding. An encoding
642 to transfer strings in this form (e.g. to write them to a file) would
645 pack('L',map(chr($_),split(//,$string))); # native
647 pack('V',map(chr($_),split(//,$string))); # little-endian
649 pack('N',map(chr($_),split(//,$string))); # big-endian
651 depending on the endian required.
653 No UTF-32 encodings are implemented yet.
655 Both UCS-2 and UCS-4 style encodings can have "byte order marks" by
656 representing the code point 0xFFFE as the very first thing in a file.
658 =head2 Listing available encodings
660 use Encode qw(encodings);
663 Returns a list of the canonical names of the available encodings.
665 =head2 Defining Aliases
667 use Encode qw(define_alias);
668 define_alias( newName => ENCODING);
670 Allows newName to be used as am alias for ENCODING. ENCODING may be
671 either the name of an encoding or and encoding object (as above).
673 Currently I<newName> can be specified in the following ways:
677 =item As a simple string.
679 =item As a qr// compiled regular expression, e.g.:
681 define_alias( qr/^iso8859-(\d+)$/i => '"iso-8859-$1"' );
683 In this case if I<ENCODING> is not a reference it is C<eval>-ed to
684 allow C<$1> etc. to be subsituted. The example is one way to names as
685 used in X11 font names to alias the MIME names for the iso-8859-*
688 =item As a code reference, e.g.:
690 define_alias( sub { return /^iso8859-(\d+)$/i ? "iso-8859-$1" : undef } , '');
692 In this case C<$_> will be set to the name that is being looked up and
693 I<ENCODING> is passed to the sub as its first argument. The example
694 is another way to names as used in X11 font names to alias the MIME
695 names for the iso-8859-* family.
699 =head2 Defining Encodings
701 use Encode qw(define_alias);
702 define_encoding( $object, 'canonicalName' [,alias...]);
704 Causes I<canonicalName> to be associated with I<$object>. The object
705 should provide the interface described in L</"IMPLEMENTATION CLASSES">
706 below. If more than two arguments are provided then additional
707 arguments are taken as aliases for I<$object> as for C<define_alias>.
709 =head1 Encoding and IO
711 It is very common to want to do encoding transformations when
712 reading or writing files, network connections, pipes etc.
713 If Perl is configured to use the new 'perlio' IO system then
714 C<Encode> provides a "layer" (See L<perliol>) which can transform
715 data as it is read or written.
717 open(my $ilyad,'>:encoding(iso-8859-7)','ilyad.greek');
720 In addition the new IO system can also be configured to read/write
721 UTF-8 encoded characters (as noted above this is efficient):
723 open(my $fh,'>:utf8','anything');
724 print $fh "Any \x{0021} string \N{SMILEY FACE}\n";
726 Either of the above forms of "layer" specifications can be made the default
727 for a lexical scope with the C<use open ...> pragma. See L<open>.
729 Once a handle is open is layers can be altered using C<binmode>.
731 Without any such configuration, or if Perl itself is built using
732 system's own IO, then write operations assume that file handle accepts
733 only I<bytes> and will C<die> if a character larger than 255 is
734 written to the handle. When reading, each octet from the handle
735 becomes a byte-in-a-character. Note that this default is the same
736 behaviour as bytes-only languages (including Perl before v5.6) would
737 have, and is sufficient to handle native 8-bit encodings
738 e.g. iso-8859-1, EBCDIC etc. and any legacy mechanisms for handling
739 other encodings and binary data.
741 In other cases it is the programs responsibility to transform
742 characters into bytes using the API above before doing writes, and to
743 transform the bytes read from a handle into characters before doing
744 "character operations" (e.g. C<lc>, C</\W+/>, ...).
746 =head1 Encode and PerlIO
748 The PerlIO layer (new since Perl 5.7) can be used to automatically
749 convert the data being read in or written out to be converted from
750 some encoding into Perl's internal encoding or from Perl's internal
751 encoding into some other encoding.
755 open(my $f, "<:encoding(cp1252)")
757 open(my $g, ">:encoding(iso-8859-1)")
759 You can also use PerlIO to convert larger amounts of data you don't
760 want to bring into memory. For example to convert between ISO 8859-1
761 (Latin 1) and UTF-8 (or UTF-EBCDIC in EBCDIC machines):
763 open(F, "<:encoding(iso-8859-1)", "data.txt") or die $!;
764 open(G, ">:utf8", "data.utf") or die $!;
765 while (<F>) { print G }
767 # Could also do "print G <F>" but that would pull
768 # the whole file into memory just to write it out again.
770 See L<PerlIO> for more information.
772 =head1 Encoding How to ...
778 =item * IO with mixed content (faking iso-2020-*)
780 =item * MIME's Content-Length:
782 =item * UTF-8 strings in binary data.
784 =item * Perl/Encode wrappers on non-Unicode XS modules.
788 =head1 Messing with Perl's Internals
790 The following API uses parts of Perl's internals in the current
791 implementation. As such they are efficient, but may change.
795 =item * is_utf8(STRING [, CHECK])
797 [INTERNAL] Test whether the UTF-8 flag is turned on in the STRING.
798 If CHECK is true, also checks the data in STRING for being well-formed
799 UTF-8. Returns true if successful, false otherwise.
801 =item * valid_utf8(STRING)
803 [INTERNAL] Test whether STRING is in a consistent state. Will return
804 true if string is held as bytes, or is well-formed UTF-8 and has the
805 UTF-8 flag on. Main reason for this routine is to allow Perl's
806 testsuite to check that operations have left strings in a consistent
813 [INTERNAL] Turn on the UTF-8 flag in STRING. The data in STRING is
814 B<not> checked for being well-formed UTF-8. Do not use unless you
815 B<know> that the STRING is well-formed UTF-8. Returns the previous
816 state of the UTF-8 flag (so please don't test the return value as
817 I<not> success or failure), or C<undef> if STRING is not a string.
823 [INTERNAL] Turn off the UTF-8 flag in STRING. Do not use frivolously.
824 Returns the previous state of the UTF-8 flag (so please don't test the
825 return value as I<not> success or failure), or C<undef> if STRING is
830 =head1 IMPLEMENTATION CLASSES
832 As mentioned above encodings are (in the current implementation at least)
833 defined by objects. The mapping of encoding name to object is via the
836 The values of the hash can currently be either strings or objects.
837 The string form may go away in the future. The string form occurs
838 when C<encodings()> has scanned C<@INC> for loadable encodings but has
839 not actually loaded the encoding in question. This is because the
840 current "loading" process is all Perl and a bit slow.
842 Once an encoding is loaded then value of the hash is object which
843 implements the encoding. The object should provide the following
850 Should return the string representing the canonical name of the encoding.
852 =item -E<gt>new_sequence
854 This is a placeholder for encodings with state. It should return an
855 object which implements this interface, all current implementations
856 return the original object.
858 =item -E<gt>encode($string,$check)
860 Should return the octet sequence representing I<$string>. If I<$check>
861 is true it should modify I<$string> in place to remove the converted
862 part (i.e. the whole string unless there is an error). If an error
863 occurs it should return the octet sequence for the fragment of string
864 that has been converted, and modify $string in-place to remove the
865 converted part leaving it starting with the problem fragment.
867 If check is is false then C<encode> should make a "best effort" to
868 convert the string - for example by using a replacement character.
870 =item -E<gt>decode($octets,$check)
872 Should return the string that I<$octets> represents. If I<$check> is
873 true it should modify I<$octets> in place to remove the converted part
874 (i.e. the whole sequence unless there is an error). If an error
875 occurs it should return the fragment of string that has been
876 converted, and modify $octets in-place to remove the converted part
877 leaving it starting with the problem fragment.
879 If check is is false then C<decode> should make a "best effort" to
880 convert the string - for example by using Unicode's "\x{FFFD}" as a
881 replacement character.
885 It should be noted that the check behaviour is different from the
886 outer public API. The logic is that the "unchecked" case is useful
887 when encoding is part of a stream which may be reporting errors
888 (e.g. STDERR). In such cases it is desirable to get everything
889 through somehow without causing additional errors which obscure the
890 original one. Also the encoding is best placed to know what the
891 correct replacement character is, so if that is the desired behaviour
892 then letting low level code do it is the most efficient.
894 In contrast if check is true, the scheme above allows the encoding to
895 do as much as it can and tell layer above how much that was. What is
896 lacking at present is a mechanism to report what went wrong. The most
897 likely interface will be an additional method call to the object, or
898 perhaps (to avoid forcing per-stream objects on otherwise stateless
899 encodings) and additional parameter.
901 It is also highly desirable that encoding classes inherit from
902 C<Encode::Encoding> as a base class. This allows that class to define
903 additional behaviour for all encoding objects. For example built in
904 Unicode, UCS-2 and UTF-8 classes use :
906 package Encode::MyEncoding;
907 use base qw(Encode::Encoding);
909 __PACKAGE__->Define(qw(myCanonical myAlias));
911 To create an object with bless {Name => ...},$class, and call
912 define_encoding. They inherit their C<name> method from
915 =head2 Compiled Encodings
917 F<Encode.xs> provides a class C<Encode::XS> which provides the
918 interface described above. It calls a generic octet-sequence to
919 octet-sequence "engine" that is driven by tables (defined in
920 F<encengine.c>). The same engine is used for both encode and
921 decode. C<Encode:XS>'s C<encode> forces Perl's characters to their
922 UTF-8 form and then treats them as just another multibyte
923 encoding. C<Encode:XS>'s C<decode> transforms the sequence and then
924 turns the UTF-8-ness flag as that is the form that the tables are
925 defined to produce. For details of the engine see the comments in
928 The tables are produced by the Perl script F<compile> (the name needs
929 to change so we can eventually install it somewhere). F<compile> can
930 currently read two formats:
936 This is a coined format used by Tcl. It is documented in
937 Encode/EncodeFormat.pod.
941 This is the semi-standard format used by IBM's ICU package.
945 F<compile> can write the following forms:
951 See above - the F<Encode/*.ucm> files provided with the distribution have
952 been created from the original Tcl .enc files using this approach.
956 Produces tables as C data structures - this is used to build in encodings
957 into F<Encode.so>/F<Encode.dll>.
961 In theory this allows encodings to be stand-alone loadable Perl
962 extensions. The process has not yet been tested. The plan is to use
963 this approach for large East Asian encodings.
967 The set of encodings built-in to F<Encode.so>/F<Encode.dll> is
968 determined by F<Makefile.PL>. The current set is as follows:
972 =item ascii and iso-8859-*
974 That is all the common 8-bit "western" encodings.
976 =item IBM-1047 and two other variants of EBCDIC.
978 These are the same variants that are supported by EBCDIC Perl as
979 "native" encodings. They are included to prove "reversibility" of
980 some constructs in EBCDIC Perl.
982 =item symbol and dingbats as used by Tk on X11.
984 (The reason Encode got started was to support Perl/Tk.)
988 That set is rather ad hoc and has been driven by the needs of the
989 tests rather than the needs of typical applications. It is likely
994 L<perlunicode>, L<perlebcdic>, L<perlfunc/open>, L<PerlIO>