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
45 # 0 1 2 3 4 5 6 7 8 9 10
46 our @latin2iso_num = ( 0, 1, 2, 3, 4, 9, 10, 13, 14, 15, 16 );
52 return keys %encoding;
59 unless (exists $alias{$_})
61 for (my $i=0; $i < @alias; $i += 2)
63 my $alias = $alias[$i];
64 my $val = $alias[$i+1];
66 if (ref($alias) eq 'Regexp' && $_ =~ $alias)
70 elsif (ref($alias) eq 'CODE')
72 $new = &{$alias}($val)
74 elsif (lc($_) eq lc($alias))
80 next if $new eq $_; # avoid (direct) recursion on bugs
81 my $enc = (ref($new)) ? $new : find_encoding($new);
97 my ($alias,$name) = splice(@_,0,2);
98 push(@alias, $alias => $name);
102 # Allow variants of iso-8859-1 etc.
103 define_alias( qr/^iso[-_]?(\d+)[-_](\d+)$/i => '"iso-$1-$2"' );
105 # Allow latin-1 style names as well
106 define_alias( qr/^latin[-_]?(\d+)$/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,$name) = @_;
139 if (ref($name) && $name->can('new_sequence'))
143 if (exists $encoding{$name})
145 return $encoding{$name};
149 return $class->findAlias($name);
156 return __PACKAGE__->getEncoding($name);
161 my ($name,$string,$check) = @_;
162 my $enc = find_encoding($name);
163 croak("Unknown encoding '$name'") unless defined $enc;
164 my $octets = $enc->encode($string,$check);
165 return undef if ($check && length($string));
171 my ($name,$octets,$check) = @_;
172 my $enc = find_encoding($name);
173 croak("Unknown encoding '$name'") unless defined $enc;
174 my $string = $enc->decode($octets,$check);
175 return undef if ($check && length($octets));
181 my ($string,$from,$to,$check) = @_;
182 my $f = find_encoding($from);
183 croak("Unknown encoding '$from'") unless defined $f;
184 my $t = find_encoding($to);
185 croak("Unknown encoding '$to'") unless defined $t;
186 my $uni = $f->decode($string,$check);
187 return undef if ($check && length($string));
188 $string = $t->encode($uni,$check);
189 return undef if ($check && length($uni));
190 return length($_[0] = $string);
203 return undef unless utf8::decode($str);
207 package Encode::Encoding;
208 # Base class for classes which implement encodings
213 my $canonical = shift;
214 $obj = bless { Name => $canonical },$obj unless ref $obj;
215 # warn "$canonical => $obj\n";
216 Encode::define_encoding($obj, $canonical, @_);
219 sub name { shift->{'Name'} }
221 # Temporary legacy methods
222 sub toUnicode { shift->decode(@_) }
223 sub fromUnicode { shift->encode(@_) }
225 sub new_sequence { return $_[0] }
228 use base 'Encode::Encoding';
230 package Encode::Internal;
231 use base 'Encode::Encoding';
233 # Dummy package that provides the encode interface but leaves data
234 # as UTF-X encoded. It is here so that from_to() works.
236 __PACKAGE__->Define('Internal');
238 Encode::define_alias( 'Unicode' => 'Internal' ) if ord('A') == 65;
242 my ($obj,$str,$chk) = @_;
250 package Encoding::Unicode;
251 use base 'Encode::Encoding';
253 __PACKAGE__->Define('Unicode') unless ord('A') == 65;
257 my ($obj,$str,$chk) = @_;
259 for (my $i = 0; $i < length($str); $i++)
261 $res .= chr(utf8::unicode_to_native(ord(substr($str,$i,1))));
269 my ($obj,$str,$chk) = @_;
271 for (my $i = 0; $i < length($str); $i++)
273 $res .= chr(utf8::native_to_unicode(ord(substr($str,$i,1))));
280 package Encode::utf8;
281 use base 'Encode::Encoding';
282 # package to allow long-hand
283 # $octets = encode( utf8 => $string );
286 __PACKAGE__->Define(qw(UTF-8 utf8));
290 my ($obj,$octets,$chk) = @_;
291 my $str = Encode::decode_utf8($octets);
302 my ($obj,$string,$chk) = @_;
303 my $octets = Encode::encode_utf8($string);
308 package Encode::iso10646_1;
309 use base 'Encode::Encoding';
310 # Encoding is 16-bit network order Unicode (no surogates)
311 # Used for X font encodings
313 __PACKAGE__->Define(qw(UCS-2 iso-10646-1));
317 my ($obj,$str,$chk) = @_;
321 my $code = unpack('n',substr($str,0,2,'')) & 0xffff;
324 $_[1] = $str if $chk;
331 my ($obj,$uni,$chk) = @_;
335 my $ch = substr($uni,0,1,'');
342 $str .= pack('n',$x);
344 $_[1] = $uni if $chk;
348 # switch back to Encode package in case we ever add AutoLoader
357 Encode - character encodings
365 The C<Encode> module provides the interfaces between Perl's strings
366 and the rest of the system. Perl strings are sequences of B<characters>.
368 The repertoire of characters that Perl can represent is at least that
369 defined by the Unicode Consortium. On most platforms the ordinal
370 values of the characters (as returned by C<ord(ch)>) is the "Unicode
371 codepoint" for the character (the exceptions are those platforms where
372 the legacy encoding is some variant of EBCDIC rather than a super-set
373 of ASCII - see L<perlebcdic>).
375 Traditionaly computer data has been moved around in 8-bit chunks
376 often called "bytes". These chunks are also known as "octets" in
377 networking standards. Perl is widely used to manipulate data of
378 many types - not only strings of characters representing human or
379 computer languages but also "binary" data being the machines representation
380 of numbers, pixels in an image - or just about anything.
382 When Perl is processing "binary data" the programmer wants Perl to process
383 "sequences of bytes". This is not a problem for Perl - as a byte has 256
384 possible values it easily fits in Perl's much larger "logical character".
392 I<character>: a character in the range 0..(2**32-1) (or more).
393 (What Perl's strings are made of.)
397 I<byte>: a character in the range 0..255
398 (A special case of a Perl character.)
402 I<octet>: 8 bits of data, with ordinal values 0..255
403 (Term for bytes passed to or from a non-Perl context, e.g. disk file.)
407 The marker [INTERNAL] marks Internal Implementation Details, in
408 general meant only for those who think they know what they are doing,
409 and such details may change in future releases.
413 =head2 Characteristics of an Encoding
415 An encoding has a "repertoire" of characters that it can represent,
416 and for each representable character there is at least one sequence of
417 octets that represents it.
419 =head2 Types of Encodings
421 Encodings can be divided into the following types:
425 =item * Fixed length 8-bit (or less) encodings.
427 Each character is a single octet so may have a repertoire of up to
428 256 characters. ASCII and iso-8859-* are typical examples.
430 =item * Fixed length 16-bit encodings
432 Each character is two octets so may have a repertoire of up to
433 65 536 characters. Unicode's UCS-2 is an example. Also used for
434 encodings for East Asian languages.
436 =item * Fixed length 32-bit encodings.
438 Not really very "encoded" encodings. The Unicode code points
439 are just represented as 4-octet integers. None the less because
440 different architectures use different representations of integers
441 (so called "endian") there at least two disctinct encodings.
443 =item * Multi-byte encodings
445 The number of octets needed to represent a character varies.
446 UTF-8 is a particularly complex but regular case of a multi-byte
447 encoding. Several East Asian countries use a multi-byte encoding
448 where 1-octet is used to cover western roman characters and Asian
449 characters get 2-octets.
450 (UTF-16 is strictly a multi-byte encoding taking either 2 or 4 octets
451 to represent a Unicode code point.)
453 =item * "Escape" encodings.
455 These encodings embed "escape sequences" into the octet sequence
456 which describe how the following octets are to be interpreted.
457 The iso-2022-* family is typical. Following the escape sequence
458 octets are encoded by an "embedded" encoding (which will be one
459 of the above types) until another escape sequence switches to
460 a different "embedded" encoding.
462 These schemes are very flexible and can handle mixed languages but are
463 very complex to process (and have state). No escape encodings are
464 implemented for Perl yet.
468 =head2 Specifying Encodings
470 Encodings can be specified to the API described below in two ways:
476 Encoding names are strings with characters taken from a restricted
477 repertoire. See L</"Encoding Names">.
479 =item 2. As an object
481 Encoding objects are returned by C<find_encoding($name)>.
485 =head2 Encoding Names
487 Encoding names are case insensitive. White space in names is ignored.
488 In addition an encoding may have aliases. Each encoding has one
489 "canonical" name. The "canonical" name is chosen from the names of
490 the encoding by picking the first in the following sequence:
494 =item * The MIME name as defined in IETF RFC-XXXX.
496 =item * The name in the IANA registry.
498 =item * The name used by the the organization that defined it.
502 Because of all the alias issues, and because in the general case
503 encodings have state C<Encode> uses the encoding object internally
504 once an operation is in progress.
506 =head1 PERL ENCODING API
508 =head2 Generic Encoding Interface
514 $bytes = encode(ENCODING, $string[, CHECK])
516 Encodes string from Perl's internal form into I<ENCODING> and returns
517 a sequence of octets. For CHECK see L</"Handling Malformed Data">.
521 $string = decode(ENCODING, $bytes[, CHECK])
523 Decode sequence of octets assumed to be in I<ENCODING> into Perl's
524 internal form and returns the resulting string. For CHECK see
525 L</"Handling Malformed Data">.
529 from_to($string, FROM_ENCODING, TO_ENCODING[, CHECK])
531 Convert B<in-place> the data between two encodings. How did the data
532 in $string originally get to be in FROM_ENCODING? Either using
533 encode() or through PerlIO: See L</"Encoding and IO">. For CHECK
534 see L</"Handling Malformed Data">.
536 For example to convert ISO 8859-1 data to UTF-8:
538 from_to($data, "iso-8859-1", "utf-8");
540 and to convert it back:
542 from_to($data, "utf-8", "iso-8859-1");
544 Note that because the conversion happens in place, the data to be
545 converted cannot be a string constant, it must be a scalar variable.
549 =head2 Handling Malformed Data
551 If CHECK is not set, C<undef> is returned. If the data is supposed to
552 be UTF-8, an optional lexical warning (category utf8) is given. If
553 CHECK is true but not a code reference, dies.
555 It would desirable to have a way to indicate that transform should use
556 the encodings "replacement character" - no such mechanism is defined yet.
558 It is also planned to allow I<CHECK> to be a code reference.
560 This is not yet implemented as there are design issues with what its
561 arguments should be and how it returns its results.
567 Passed remaining fragment of string being processed.
568 Modifies it in place to remove bytes/characters it can understand
569 and returns a string used to represent them.
573 my $ch = substr($_[0],0,1,'');
574 return sprintf("\x{%02X}",ord($ch);
577 This scheme is close to how underlying C code for Encode works, but gives
578 the fixup routine very little context.
582 Passed original string, and an index into it of the problem area, and
583 output string so far. Appends what it will to output string and
584 returns new index into original string. For example:
587 # my ($s,$i,$d) = @_;
588 my $ch = substr($_[0],$_[1],1);
589 $_[2] .= sprintf("\x{%02X}",ord($ch);
593 This scheme gives maximal control to the fixup routine but is more
594 complicated to code, and may need internals of Encode to be tweaked to
595 keep original string intact.
601 Multiple return values rather than in-place modifications.
603 Index into the string could be pos($str) allowing s/\G...//.
609 The Unicode consortium defines the UTF-8 standard as a way of encoding
610 the entire Unicode repertiore as sequences of octets. This encoding is
611 expected to become very widespread. Perl can use this form internaly
612 to represent strings, so conversions to and from this form are
613 particularly efficient (as octets in memory do not have to change,
614 just the meta-data that tells Perl how to treat them).
620 $bytes = encode_utf8($string);
622 The characters that comprise string are encoded in Perl's superset of UTF-8
623 and the resulting octets returned as a sequence of bytes. All possible
624 characters have a UTF-8 representation so this function cannot fail.
628 $string = decode_utf8($bytes [,CHECK]);
630 The sequence of octets represented by $bytes is decoded from UTF-8
631 into a sequence of logical characters. Not all sequences of octets
632 form valid UTF-8 encodings, so it is possible for this call to fail.
633 For CHECK see L</"Handling Malformed Data">.
637 =head2 Other Encodings of Unicode
639 UTF-16 is similar to UCS-2, 16 bit or 2-byte chunks. UCS-2 can only
640 represent 0..0xFFFF, while UTF-16 has a "surrogate pair" scheme which
641 allows it to cover the whole Unicode range.
643 Encode implements big-endian UCS-2 aliased to "iso-10646-1" as that
644 happens to be the name used by that representation when used with X11
647 UTF-32 or UCS-4 is 32-bit or 4-byte chunks. Perl's logical characters
648 can be considered as being in this form without encoding. An encoding
649 to transfer strings in this form (e.g. to write them to a file) would
652 pack('L',map(chr($_),split(//,$string))); # native
654 pack('V',map(chr($_),split(//,$string))); # little-endian
656 pack('N',map(chr($_),split(//,$string))); # big-endian
658 depending on the endian required.
660 No UTF-32 encodings are implemented yet.
662 Both UCS-2 and UCS-4 style encodings can have "byte order marks" by
663 representing the code point 0xFFFE as the very first thing in a file.
665 =head2 Listing available encodings
667 use Encode qw(encodings);
670 Returns a list of the canonical names of the available encodings.
672 =head2 Defining Aliases
674 use Encode qw(define_alias);
675 define_alias( newName => ENCODING);
677 Allows newName to be used as am alias for ENCODING. ENCODING may be
678 either the name of an encoding or and encoding object (as above).
680 Currently I<newName> can be specified in the following ways:
684 =item As a simple string.
686 =item As a qr// compiled regular expression, e.g.:
688 define_alias( qr/^iso8859-(\d+)$/i => '"iso-8859-$1"' );
690 In this case if I<ENCODING> is not a reference it is C<eval>-ed to
691 allow C<$1> etc. to be subsituted. The example is one way to names as
692 used in X11 font names to alias the MIME names for the iso-8859-*
695 =item As a code reference, e.g.:
697 define_alias( sub { return /^iso8859-(\d+)$/i ? "iso-8859-$1" : undef } , '');
699 In this case C<$_> will be set to the name that is being looked up and
700 I<ENCODING> is passed to the sub as its first argument. The example
701 is another way to names as used in X11 font names to alias the MIME
702 names for the iso-8859-* family.
706 =head2 Defining Encodings
708 use Encode qw(define_alias);
709 define_encoding( $object, 'canonicalName' [,alias...]);
711 Causes I<canonicalName> to be associated with I<$object>. The object
712 should provide the interface described in L</"IMPLEMENTATION CLASSES">
713 below. If more than two arguments are provided then additional
714 arguments are taken as aliases for I<$object> as for C<define_alias>.
716 =head1 Encoding and IO
718 It is very common to want to do encoding transformations when
719 reading or writing files, network connections, pipes etc.
720 If Perl is configured to use the new 'perlio' IO system then
721 C<Encode> provides a "layer" (See L<perliol>) which can transform
722 data as it is read or written.
724 open(my $ilyad,'>:encoding(iso-8859-7)','ilyad.greek');
727 In addition the new IO system can also be configured to read/write
728 UTF-8 encoded characters (as noted above this is efficient):
730 open(my $fh,'>:utf8','anything');
731 print $fh "Any \x{0021} string \N{SMILEY FACE}\n";
733 Either of the above forms of "layer" specifications can be made the default
734 for a lexical scope with the C<use open ...> pragma. See L<open>.
736 Once a handle is open is layers can be altered using C<binmode>.
738 Without any such configuration, or if Perl itself is built using
739 system's own IO, then write operations assume that file handle accepts
740 only I<bytes> and will C<die> if a character larger than 255 is
741 written to the handle. When reading, each octet from the handle
742 becomes a byte-in-a-character. Note that this default is the same
743 behaviour as bytes-only languages (including Perl before v5.6) would
744 have, and is sufficient to handle native 8-bit encodings
745 e.g. iso-8859-1, EBCDIC etc. and any legacy mechanisms for handling
746 other encodings and binary data.
748 In other cases it is the programs responsibility to transform
749 characters into bytes using the API above before doing writes, and to
750 transform the bytes read from a handle into characters before doing
751 "character operations" (e.g. C<lc>, C</\W+/>, ...).
753 You can also use PerlIO to convert larger amounts of data you don't
754 want to bring into memory. For example to convert between ISO 8859-1
755 (Latin 1) and UTF-8 (or UTF-EBCDIC in EBCDIC machines):
757 open(F, "<:encoding(iso-8859-1)", "data.txt") or die $!;
758 open(G, ">:utf8", "data.utf") or die $!;
759 while (<F>) { print G }
761 # Could also do "print G <F>" but that would pull
762 # the whole file into memory just to write it out again.
766 open(my $f, "<:encoding(cp1252)")
767 open(my $g, ">:encoding(iso-8859-2)")
768 open(my $h, ">:encoding(latin9)") # iso-8859-15
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>