2 # $Id: Encode.pm,v 2.75 2015/06/30 09:57:15 dankogai Exp $
7 our $VERSION = sprintf "%d.%02d", q$Revision: 2.75 $ =~ /(\d+)/g;
8 use constant DEBUG => !!$ENV{PERL_ENCODE_DEBUG};
10 XSLoader::load( __PACKAGE__, $VERSION );
12 use Exporter 5.57 'import';
14 # Public, encouraged API is exported by default
17 decode decode_utf8 encode encode_utf8 str2bytes bytes2str
18 encodings find_encoding clone_encoding
21 DIE_ON_ERR WARN_ON_ERR RETURN_ON_ERR LEAVE_SRC
22 PERLQQ HTMLCREF XMLCREF STOP_AT_PARTIAL
25 FB_DEFAULT FB_CROAK FB_QUIET FB_WARN
26 FB_PERLQQ FB_HTMLCREF FB_XMLCREF
30 _utf8_off _utf8_on define_encoding from_to is_16bit is_8bit
31 is_utf8 perlio_ok resolve_alias utf8_downgrade utf8_upgrade
33 @FB_FLAGS, @FB_CONSTS,
37 all => [ @EXPORT, @EXPORT_OK ],
38 default => [ @EXPORT ],
39 fallbacks => [ @FB_CONSTS ],
40 fallback_all => [ @FB_CONSTS, @FB_FLAGS ],
43 # Documentation moved after __END__ for speed - NI-S
45 our $ON_EBCDIC = ( ord("A") == 193 );
49 # Make a %Encoding package variable to allow a certain amount of cheating
52 require Encode::Config;
54 # https://bugzilla.redhat.com/show_bug.cgi?id=435505#c2
55 # to find why sig handlers inside eval{} are disabled.
59 require Encode::ConfigLocal;
64 my $arg = $_[1] || '';
65 if ( $arg eq ":all" ) {
66 %enc = ( %Encoding, %ExtModule );
70 for my $mod ( map { m/::/ ? $_ : "Encode::$_" } @_ ) {
72 for my $enc ( keys %ExtModule ) {
73 $ExtModule{$enc} eq $mod and $enc{$enc} = $mod;
77 return sort { lc $a cmp lc $b }
78 grep { !/^(?:Internal|Unicode|Guess)$/o } keys %enc;
82 my $obj = ref( $_[0] ) ? $_[0] : find_encoding( $_[0] );
83 $obj->can("perlio_ok") and return $obj->perlio_ok();
84 return 0; # safety net
90 $Encoding{$name} = $obj;
92 define_alias( $lc => $obj ) unless $lc eq $name;
95 define_alias( $alias, $obj );
101 my ( $class, $name, $skip_external ) = @_;
103 $name =~ s/\s+//g; # https://rt.cpan.org/Ticket/Display.html?id=65796
105 ref($name) && $name->can('renew') and return $name;
106 exists $Encoding{$name} and return $Encoding{$name};
108 exists $Encoding{$lc} and return $Encoding{$lc};
110 my $oc = $class->find_alias($name);
111 defined($oc) and return $oc;
112 $lc ne $name and $oc = $class->find_alias($lc);
113 defined($oc) and return $oc;
115 unless ($skip_external) {
116 if ( my $mod = $ExtModule{$name} || $ExtModule{$lc} ) {
119 eval { require $mod; };
120 exists $Encoding{$name} and return $Encoding{$name};
126 sub find_encoding($;$) {
127 my ( $name, $skip_external ) = @_;
128 return __PACKAGE__->getEncoding( $name, $skip_external );
131 sub resolve_alias($) {
132 my $obj = find_encoding(shift);
133 defined $obj and return $obj->name;
137 sub clone_encoding($) {
138 my $obj = find_encoding(shift);
140 eval { require Storable };
142 return Storable::dclone($obj);
146 my ( $name, $string, $check ) = @_;
147 return undef unless defined $string;
148 $string .= ''; # stringify;
150 unless ( defined $name ) {
152 Carp::croak("Encoding name should not be undef");
154 my $enc = find_encoding($name);
155 unless ( defined $enc ) {
157 Carp::croak("Unknown encoding '$name'");
159 # For Unicode, warnings need to be caught and re-issued at this level
160 # so that callers can disable utf8 warnings lexically.
162 if ( ref($enc) eq 'Encode::Unicode' ) {
165 local $SIG{__WARN__} = sub { $warn = shift };
166 $octets = $enc->encode( $string, $check );
168 warnings::warnif('utf8', $warn) if length $warn;
171 $octets = $enc->encode( $string, $check );
173 $_[1] = $string if $check and !ref $check and !( $check & LEAVE_SRC() );
176 *str2bytes = \&encode;
179 my ( $name, $octets, $check ) = @_;
180 return undef unless defined $octets;
183 my $enc = find_encoding($name);
184 unless ( defined $enc ) {
186 Carp::croak("Unknown encoding '$name'");
188 # For Unicode, warnings need to be caught and re-issued at this level
189 # so that callers can disable utf8 warnings lexically.
191 if ( ref($enc) eq 'Encode::Unicode' ) {
194 local $SIG{__WARN__} = sub { $warn = shift };
195 $string = $enc->decode( $octets, $check );
197 warnings::warnif('utf8', $warn) if length $warn;
200 $string = $enc->decode( $octets, $check );
202 $_[1] = $octets if $check and !ref $check and !( $check & LEAVE_SRC() );
205 *bytes2str = \&decode;
208 my ( $string, $from, $to, $check ) = @_;
209 return undef unless defined $string;
211 my $f = find_encoding($from);
212 unless ( defined $f ) {
214 Carp::croak("Unknown encoding '$from'");
216 my $t = find_encoding($to);
217 unless ( defined $t ) {
219 Carp::croak("Unknown encoding '$to'");
221 my $uni = $f->decode($string);
222 $_[0] = $string = $t->encode( $uni, $check );
223 return undef if ( $check && length($uni) );
224 return defined( $_[0] ) ? length($string) : undef;
235 sub decode_utf8($;$) {
236 my ( $octets, $check ) = @_;
237 return undef unless defined $octets;
240 $utf8enc ||= find_encoding('utf8');
241 my $string = $utf8enc->decode( $octets, $check );
242 $_[0] = $octets if $check and !ref $check and !( $check & LEAVE_SRC() );
246 # sub decode_utf8($;$) {
247 # my ( $str, $check ) = @_;
248 # return $str if is_utf8($str);
250 # return decode( "utf8", $str, $check );
253 # return decode( "utf8", $str );
258 predefine_encodings(1);
261 # This is to restore %Encoding if really needed;
264 sub predefine_encodings {
265 require Encode::Encoding;
266 no warnings 'redefine';
270 # was in Encode::UTF_EBCDIC
271 package Encode::UTF_EBCDIC;
272 push @Encode::UTF_EBCDIC::ISA, 'Encode::Encoding';
274 my ( undef, $str, $chk ) = @_;
276 for ( my $i = 0 ; $i < length($str) ; $i++ ) {
279 utf8::unicode_to_native( ord( substr( $str, $i, 1 ) ) )
286 my ( undef, $str, $chk ) = @_;
288 for ( my $i = 0 ; $i < length($str) ; $i++ ) {
291 utf8::native_to_unicode( ord( substr( $str, $i, 1 ) ) )
297 $Encode::Encoding{Unicode} =
298 bless { Name => "UTF_EBCDIC" } => "Encode::UTF_EBCDIC";
302 package Encode::Internal;
303 push @Encode::Internal::ISA, 'Encode::Encoding';
305 my ( undef, $str, $chk ) = @_;
311 $Encode::Encoding{Unicode} =
312 bless { Name => "Internal" } => "Encode::Internal";
315 # https://rt.cpan.org/Public/Bug/Display.html?id=103253
317 push @Encode::XS::ISA, 'Encode::Encoding';
321 # was in Encode::utf8
322 package Encode::utf8;
323 push @Encode::utf8::ISA, 'Encode::Encoding';
327 Encode::DEBUG and warn __PACKAGE__, " XS on";
328 *decode = \&decode_xs;
329 *encode = \&encode_xs;
332 Encode::DEBUG and warn __PACKAGE__, " XS off";
334 my ( undef, $octets, $chk ) = @_;
335 my $str = Encode::decode_utf8($octets);
336 if ( defined $str ) {
343 my ( undef, $string, $chk ) = @_;
344 my $octets = Encode::encode_utf8($string);
349 *cat_decode = sub { # ($obj, $dst, $src, $pos, $trm, $chk)
350 # currently ignores $chk
351 my ( undef, undef, undef, $pos, $trm ) = @_;
352 my ( $rdst, $rsrc, $rpos ) = \@_[ 1, 2, 3 ];
354 if ( ( my $npos = index( $$rsrc, $trm, $pos ) ) >= 0 ) {
356 substr( $$rsrc, $pos, $npos - $pos + length($trm) );
357 $$rpos = $npos + length($trm);
360 $$rdst .= substr( $$rsrc, $pos );
361 $$rpos = length($$rsrc);
364 $Encode::Encoding{utf8} =
365 bless { Name => "utf8" } => "Encode::utf8";
366 $Encode::Encoding{"utf-8-strict"} =
367 bless { Name => "utf-8-strict", strict_utf8 => 1 }
378 Encode - character encodings in Perl
382 use Encode qw(decode encode);
383 $characters = decode('UTF-8', $octets, Encode::FB_CROAK);
384 $octets = encode('UTF-8', $characters, Encode::FB_CROAK);
386 =head2 Table of Contents
388 Encode consists of a collection of modules whose details are too extensive
389 to fit in one document. This one itself explains the top-level APIs
390 and general topics at a glance. For other topics and more details,
391 see the documentation for these modules:
395 =item L<Encode::Alias> - Alias definitions to encodings
397 =item L<Encode::Encoding> - Encode Implementation Base Class
399 =item L<Encode::Supported> - List of Supported Encodings
401 =item L<Encode::CN> - Simplified Chinese Encodings
403 =item L<Encode::JP> - Japanese Encodings
405 =item L<Encode::KR> - Korean Encodings
407 =item L<Encode::TW> - Traditional Chinese Encodings
413 The C<Encode> module provides the interface between Perl strings
414 and the rest of the system. Perl strings are sequences of
417 The repertoire of characters that Perl can represent is a superset of those
418 defined by the Unicode Consortium. On most platforms the ordinal
419 values of a character as returned by C<ord(I<S>)> is the I<Unicode
420 codepoint> for that character. The exceptions are platforms where
421 the legacy encoding is some variant of EBCDIC rather than a superset
422 of ASCII; see L<perlebcdic>.
424 During recent history, data is moved around a computer in 8-bit chunks,
425 often called "bytes" but also known as "octets" in standards documents.
426 Perl is widely used to manipulate data of many types: not only strings of
427 characters representing human or computer languages, but also "binary"
428 data, being the machine's representation of numbers, pixels in an image, or
431 When Perl is processing "binary data", the programmer wants Perl to
432 process "sequences of bytes". This is not a problem for Perl: because a
433 byte has 256 possible values, it easily fits in Perl's much larger
436 This document mostly explains the I<how>. L<perlunitut> and L<perlunifaq>
443 A character in the range 0 .. 2**32-1 (or more);
444 what Perl's strings are made of.
448 A character in the range 0..255;
449 a special case of a Perl character.
453 8 bits of data, with ordinal values 0..255;
454 term for bytes passed to or from a non-Perl context, such as a disk file,
455 standard I/O stream, database, command-line argument, environment variable,
458 =head1 THE PERL ENCODING API
464 $octets = encode(ENCODING, STRING[, CHECK])
466 Encodes the scalar value I<STRING> from Perl's internal form into
467 I<ENCODING> and returns a sequence of octets. I<ENCODING> can be either a
468 canonical name or an alias. For encoding names and aliases, see
469 L</"Defining Aliases">. For CHECK, see L</"Handling Malformed Data">.
471 For example, to convert a string from Perl's internal format into
472 ISO-8859-1, also known as Latin1:
474 $octets = encode("iso-8859-1", $string);
476 B<CAVEAT>: When you run C<$octets = encode("utf8", $string)>, then
477 $octets I<might not be equal to> $string. Though both contain the
478 same data, the UTF8 flag for $octets is I<always> off. When you
479 encode anything, the UTF8 flag on the result is always off, even when it
480 contains a completely valid utf8 string. See L</"The UTF8 flag"> below.
482 If the $string is C<undef>, then C<undef> is returned.
486 $string = decode(ENCODING, OCTETS[, CHECK])
488 This function returns the string that results from decoding the scalar
489 value I<OCTETS>, assumed to be a sequence of octets in I<ENCODING>, into
490 Perl's internal form. As with encode(),
491 I<ENCODING> can be either a canonical name or an alias. For encoding names
492 and aliases, see L</"Defining Aliases">; for I<CHECK>, see L</"Handling
495 For example, to convert ISO-8859-1 data into a string in Perl's
498 $string = decode("iso-8859-1", $octets);
500 B<CAVEAT>: When you run C<$string = decode("utf8", $octets)>, then $string
501 I<might not be equal to> $octets. Though both contain the same data, the
502 UTF8 flag for $string is on. See L</"The UTF8 flag">
505 If the $string is C<undef>, then C<undef> is returned.
509 [$obj =] find_encoding(ENCODING)
511 Returns the I<encoding object> corresponding to I<ENCODING>. Returns
512 C<undef> if no matching I<ENCODING> is find. The returned object is
513 what does the actual encoding or decoding.
515 $utf8 = decode($name, $bytes);
520 $obj = find_encoding($name);
521 croak qq(encoding "$name" not found) unless ref $obj;
522 $obj->decode($bytes);
525 with more error checking.
527 You can therefore save time by reusing this object as follows;
529 my $enc = find_encoding("iso-8859-1");
531 my $utf8 = $enc->decode($_);
532 ... # now do something with $utf8;
535 Besides L</decode> and L</encode>, other methods are
536 available as well. For instance, C<name()> returns the canonical
537 name of the encoding object.
539 find_encoding("latin1")->name; # iso-8859-1
541 See L<Encode::Encoding> for details.
545 [$length =] from_to($octets, FROM_ENC, TO_ENC [, CHECK])
547 Converts I<in-place> data between two encodings. The data in $octets
548 must be encoded as octets and I<not> as characters in Perl's internal
549 format. For example, to convert ISO-8859-1 data into Microsoft's CP1250
552 from_to($octets, "iso-8859-1", "cp1250");
554 and to convert it back:
556 from_to($octets, "cp1250", "iso-8859-1");
558 Because the conversion happens in place, the data to be
559 converted cannot be a string constant: it must be a scalar variable.
561 C<from_to()> returns the length of the converted string in octets on success,
562 and C<undef> on error.
564 B<CAVEAT>: The following operations may look the same, but are not:
566 from_to($data, "iso-8859-1", "utf8"); #1
567 $data = decode("iso-8859-1", $data); #2
569 Both #1 and #2 make $data consist of a completely valid UTF-8 string,
570 but only #2 turns the UTF8 flag on. #1 is equivalent to:
572 $data = encode("utf8", decode("iso-8859-1", $data));
574 See L</"The UTF8 flag"> below.
578 from_to($octets, $from, $to, $check);
582 $octets = encode($to, decode($from, $octets), $check);
584 Yes, it does I<not> respect the $check during decoding. It is
585 deliberately done that way. If you need minute control, use C<decode>
586 followed by C<encode> as follows:
588 $octets = encode($to, decode($from, $octets, $check_from), $check_to);
592 $octets = encode_utf8($string);
594 Equivalent to C<$octets = encode("utf8", $string)>. The characters in
595 $string are encoded in Perl's internal format, and the result is returned
596 as a sequence of octets. Because all possible characters in Perl have a
597 (loose, not strict) UTF-8 representation, this function cannot fail.
601 $string = decode_utf8($octets [, CHECK]);
603 Equivalent to C<$string = decode("utf8", $octets [, CHECK])>.
604 The sequence of octets represented by $octets is decoded
605 from UTF-8 into a sequence of logical characters.
606 Because not all sequences of octets are valid UTF-8,
607 it is quite possible for this function to fail.
608 For CHECK, see L</"Handling Malformed Data">.
610 =head2 Listing available encodings
613 @list = Encode->encodings();
615 Returns a list of canonical names of available encodings that have already
616 been loaded. To get a list of all available encodings including those that
617 have not yet been loaded, say:
619 @all_encodings = Encode->encodings(":all");
621 Or you can give the name of a specific module:
623 @with_jp = Encode->encodings("Encode::JP");
625 When "C<::>" is not in the name, "C<Encode::>" is assumed.
627 @ebcdic = Encode->encodings("EBCDIC");
629 To find out in detail which encodings are supported by this package,
630 see L<Encode::Supported>.
632 =head2 Defining Aliases
634 To add a new alias to a given encoding, use:
638 define_alias(NEWNAME => ENCODING);
640 After that, I<NEWNAME> can be used as an alias for I<ENCODING>.
641 I<ENCODING> may be either the name of an encoding or an
644 Before you do that, first make sure the alias is nonexistent using
645 C<resolve_alias()>, which returns the canonical name thereof.
648 Encode::resolve_alias("latin1") eq "iso-8859-1" # true
649 Encode::resolve_alias("iso-8859-12") # false; nonexistent
650 Encode::resolve_alias($name) eq $name # true if $name is canonical
652 C<resolve_alias()> does not need C<use Encode::Alias>; it can be
653 imported via C<use Encode qw(resolve_alias)>.
655 See L<Encode::Alias> for details.
657 =head2 Finding IANA Character Set Registry names
659 The canonical name of a given encoding does not necessarily agree with
660 IANA Character Set Registry, commonly seen as C<< Content-Type:
661 text/plain; charset=I<WHATEVER> >>. For most cases, the canonical name
662 works, but sometimes it does not, most notably with "utf-8-strict".
664 As of C<Encode> version 2.21, a new method C<mime_name()> is therefore added.
667 my $enc = find_encoding("UTF-8");
668 warn $enc->name; # utf-8-strict
669 warn $enc->mime_name; # UTF-8
671 See also: L<Encode::Encoding>
673 =head1 Encoding via PerlIO
675 If your perl supports C<PerlIO> (which is the default), you can use a
676 C<PerlIO> layer to decode and encode directly via a filehandle. The
677 following two examples are fully identical in functionality:
679 ### Version 1 via PerlIO
680 open(INPUT, "< :encoding(shiftjis)", $infile)
681 || die "Can't open < $infile for reading: $!";
682 open(OUTPUT, "> :encoding(euc-jp)", $outfile)
683 || die "Can't open > $output for writing: $!";
684 while (<INPUT>) { # auto decodes $_
685 print OUTPUT; # auto encodes $_
687 close(INPUT) || die "can't close $infile: $!";
688 close(OUTPUT) || die "can't close $outfile: $!";
690 ### Version 2 via from_to()
691 open(INPUT, "< :raw", $infile)
692 || die "Can't open < $infile for reading: $!";
693 open(OUTPUT, "> :raw", $outfile)
694 || die "Can't open > $output for writing: $!";
697 from_to($_, "shiftjis", "euc-jp", 1); # switch encoding
698 print OUTPUT; # emit raw (but properly encoded) data
700 close(INPUT) || die "can't close $infile: $!";
701 close(OUTPUT) || die "can't close $outfile: $!";
703 In the first version above, you let the appropriate encoding layer
704 handle the conversion. In the second, you explicitly translate
705 from one encoding to the other.
707 Unfortunately, it may be that encodings are not C<PerlIO>-savvy. You can check
708 to see whether your encoding is supported by C<PerlIO> by invoking the
709 C<perlio_ok> method on it:
711 Encode::perlio_ok("hz"); # false
712 find_encoding("euc-cn")->perlio_ok; # true wherever PerlIO is available
714 use Encode qw(perlio_ok); # imported upon request
717 Fortunately, all encodings that come with C<Encode> core are C<PerlIO>-savvy
718 except for C<hz> and C<ISO-2022-kr>. For the gory details, see
719 L<Encode::Encoding> and L<Encode::PerlIO>.
721 =head1 Handling Malformed Data
723 The optional I<CHECK> argument tells C<Encode> what to do when
724 encountering malformed data. Without I<CHECK>, C<Encode::FB_DEFAULT>
727 As of version 2.12, C<Encode> supports coderef values for C<CHECK>;
730 B<NOTE:> Not all encodings support this feature.
731 Some encodings ignore the I<CHECK> argument. For example,
732 L<Encode::Unicode> ignores I<CHECK> and it always croaks on error.
734 =head2 List of I<CHECK> values
738 I<CHECK> = Encode::FB_DEFAULT ( == 0)
740 If I<CHECK> is 0, encoding and decoding replace any malformed character
741 with a I<substitution character>. When you encode, I<SUBCHAR> is used.
742 When you decode, the Unicode REPLACEMENT CHARACTER, code point U+FFFD, is
743 used. If the data is supposed to be UTF-8, an optional lexical warning of
744 warning category C<"utf8"> is given.
748 I<CHECK> = Encode::FB_CROAK ( == 1)
750 If I<CHECK> is 1, methods immediately die with an error
751 message. Therefore, when I<CHECK> is 1, you should trap
752 exceptions with C<eval{}>, unless you really want to let it C<die>.
756 I<CHECK> = Encode::FB_QUIET
758 If I<CHECK> is set to C<Encode::FB_QUIET>, encoding and decoding immediately
759 return the portion of the data that has been processed so far when an
760 error occurs. The data argument is overwritten with everything
761 after that point; that is, the unprocessed portion of the data. This is
762 handy when you have to call C<decode> repeatedly in the case where your
763 source data may contain partial multi-byte character sequences,
764 (that is, you are reading with a fixed-width buffer). Here's some sample
765 code to do exactly that:
767 my($buffer, $string) = ("", "");
768 while (read($fh, $buffer, 256, length($buffer))) {
769 $string .= decode($encoding, $buffer, Encode::FB_QUIET);
770 # $buffer now contains the unprocessed partial character
775 I<CHECK> = Encode::FB_WARN
777 This is the same as C<FB_QUIET> above, except that instead of being silent
778 on errors, it issues a warning. This is handy for when you are debugging.
780 =head3 FB_PERLQQ FB_HTMLCREF FB_XMLCREF
784 =item perlqq mode (I<CHECK> = Encode::FB_PERLQQ)
786 =item HTML charref mode (I<CHECK> = Encode::FB_HTMLCREF)
788 =item XML charref mode (I<CHECK> = Encode::FB_XMLCREF)
792 For encodings that are implemented by the C<Encode::XS> module, C<CHECK> C<==>
793 C<Encode::FB_PERLQQ> puts C<encode> and C<decode> into C<perlqq> fallback mode.
795 When you decode, C<\xI<HH>> is inserted for a malformed character, where
796 I<HH> is the hex representation of the octet that could not be decoded to
797 utf8. When you encode, C<\x{I<HHHH>}> will be inserted, where I<HHHH> is
798 the Unicode code point (in any number of hex digits) of the character that
799 cannot be found in the character repertoire of the encoding.
801 The HTML/XML character reference modes are about the same. In place of
802 C<\x{I<HHHH>}>, HTML uses C<&#I<NNN>;> where I<NNN> is a decimal number, and
803 XML uses C<&#xI<HHHH>;> where I<HHHH> is the hexadecimal number.
805 In C<Encode> 2.10 or later, C<LEAVE_SRC> is also implied.
809 These modes are all actually set via a bitmask. Here is how the C<FB_I<XXX>>
810 constants are laid out. You can import the C<FB_I<XXX>> constants via
811 C<use Encode qw(:fallbacks)>, and you can import the generic bitmask
812 constants via C<use Encode qw(:fallback_all)>.
814 FB_DEFAULT FB_CROAK FB_QUIET FB_WARN FB_PERLQQ
817 RETURN_ON_ERR 0x0004 X X
827 If the C<Encode::LEAVE_SRC> bit is I<not> set but I<CHECK> is set, then the
828 source string to encode() or decode() will be overwritten in place.
829 If you're not interested in this, then bitwise-OR it with the bitmask.
831 =head2 coderef for CHECK
833 As of C<Encode> 2.12, C<CHECK> can also be a code reference which takes the
834 ordinal value of the unmapped character as an argument and returns
835 octets that represent the fallback character. For instance:
837 $ascii = encode("ascii", $utf8, sub{ sprintf "<U+%04X>", shift });
839 Acts like C<FB_PERLQQ> but U+I<XXXX> is used instead of C<\x{I<XXXX>}>.
841 Even the fallback for C<decode> must return octets, which are
842 then decoded with the character encoding that C<decode> accepts. So for
843 example if you wish to decode octets as UTF-8, and use ISO-8859-15 as
844 a fallback for bytes that are not valid UTF-8, you could write
846 $str = decode 'UTF-8', $octets, sub {
848 from_to $tmp, 'ISO-8859-15', 'UTF-8';
852 =head1 Defining Encodings
854 To define a new encoding, use:
856 use Encode qw(define_encoding);
857 define_encoding($object, CANONICAL_NAME [, alias...]);
859 I<CANONICAL_NAME> will be associated with I<$object>. The object
860 should provide the interface described in L<Encode::Encoding>.
861 If more than two arguments are provided, additional
862 arguments are considered aliases for I<$object>.
864 See L<Encode::Encoding> for details.
868 Before the introduction of Unicode support in Perl, The C<eq> operator
869 just compared the strings represented by two scalars. Beginning with
870 Perl 5.8, C<eq> compares two strings with simultaneous consideration of
871 I<the UTF8 flag>. To explain why we made it so, I quote from page 402 of
872 I<Programming Perl, 3rd ed.>
878 Old byte-oriented programs should not spontaneously break on the old
879 byte-oriented data they used to work on.
883 Old byte-oriented programs should magically start working on the new
884 character-oriented data when appropriate.
888 Programs should run just as fast in the new character-oriented mode
889 as in the old byte-oriented mode.
893 Perl should remain one language, rather than forking into a
894 byte-oriented Perl and a character-oriented Perl.
898 When I<Programming Perl, 3rd ed.> was written, not even Perl 5.6.0 had been
899 born yet, many features documented in the book remained unimplemented for a
900 long time. Perl 5.8 corrected much of this, and the introduction of the
901 UTF8 flag is one of them. You can think of there being two fundamentally
902 different kinds of strings and string-operations in Perl: one a
903 byte-oriented mode for when the internal UTF8 flag is off, and the other a
904 character-oriented mode for when the internal UTF8 flag is on.
906 Here is how C<Encode> handles the UTF8 flag.
912 When you I<encode>, the resulting UTF8 flag is always B<off>.
916 When you I<decode>, the resulting UTF8 flag is B<on>--I<unless> you can
917 unambiguously represent data. Here is what we mean by "unambiguously".
918 After C<$utf8 = decode("foo", $octet)>,
920 When $octet is... The UTF8 flag in $utf8 is
921 ---------------------------------------------
922 In ASCII only (or EBCDIC only) OFF
924 In any other Encoding ON
925 ---------------------------------------------
927 As you see, there is one exception: in ASCII. That way you can assume
928 Goal #1. And with C<Encode>, Goal #2 is assumed but you still have to be
929 careful in the cases mentioned in the B<CAVEAT> paragraphs above.
931 This UTF8 flag is not visible in Perl scripts, exactly for the same reason
932 you cannot (or rather, you I<don't have to>) see whether a scalar contains
933 a string, an integer, or a floating-point number. But you can still peek
934 and poke these if you will. See the next section.
938 =head2 Messing with Perl's Internals
940 The following API uses parts of Perl's internals in the current
941 implementation. As such, they are efficient but may change in a future
946 is_utf8(STRING [, CHECK])
948 [INTERNAL] Tests whether the UTF8 flag is turned on in the I<STRING>.
949 If I<CHECK> is true, also checks whether I<STRING> contains well-formed
950 UTF-8. Returns true if successful, false otherwise.
952 As of Perl 5.8.1, L<utf8> also has the C<utf8::is_utf8> function.
958 [INTERNAL] Turns the I<STRING>'s internal UTF8 flag B<on>. The I<STRING>
959 is I<not> checked for containing only well-formed UTF-8. Do not use this
960 unless you I<know with absolute certainty> that the STRING holds only
961 well-formed UTF-8. Returns the previous state of the UTF8 flag (so please
962 don't treat the return value as indicating success or failure), or C<undef>
963 if I<STRING> is not a string.
965 B<NOTE>: For security reasons, this function does not work on tainted values.
971 [INTERNAL] Turns the I<STRING>'s internal UTF8 flag B<off>. Do not use
972 frivolously. Returns the previous state of the UTF8 flag, or C<undef> if
973 I<STRING> is not a string. Do not treat the return value as indicative of
974 success or failure, because that isn't what it means: it is only the
977 B<NOTE>: For security reasons, this function does not work on tainted values.
979 =head1 UTF-8 vs. utf8 vs. UTF8
981 ....We now view strings not as sequences of bytes, but as sequences
982 of numbers in the range 0 .. 2**32-1 (or in the case of 64-bit
983 computers, 0 .. 2**64-1) -- Programming Perl, 3rd ed.
985 That has historically been Perl's notion of UTF-8, as that is how UTF-8 was
986 first conceived by Ken Thompson when he invented it. However, thanks to
987 later revisions to the applicable standards, official UTF-8 is now rather
988 stricter than that. For example, its range is much narrower (0 .. 0x10_FFFF
989 to cover only 21 bits instead of 32 or 64 bits) and some sequences
990 are not allowed, like those used in surrogate pairs, the 31 non-character
991 code points 0xFDD0 .. 0xFDEF, the last two code points in I<any> plane
992 (0xI<XX>_FFFE and 0xI<XX>_FFFF), all non-shortest encodings, etc.
994 The former default in which Perl would always use a loose interpretation of
995 UTF-8 has now been overruled:
997 From: Larry Wall <larry@wall.org>
998 Date: December 04, 2004 11:51:58 JST
999 To: perl-unicode@perl.org
1000 Subject: Re: Make Encode.pm support the real UTF-8
1001 Message-Id: <20041204025158.GA28754@wall.org>
1003 On Fri, Dec 03, 2004 at 10:12:12PM +0000, Tim Bunce wrote:
1004 : I've no problem with 'utf8' being perl's unrestricted uft8 encoding,
1005 : but "UTF-8" is the name of the standard and should give the
1006 : corresponding behaviour.
1008 For what it's worth, that's how I've always kept them straight in my
1011 Also for what it's worth, Perl 6 will mostly default to strict but
1012 make it easy to switch back to lax.
1016 Got that? As of Perl 5.8.7, B<"UTF-8"> means UTF-8 in its current
1017 sense, which is conservative and strict and security-conscious, whereas
1018 B<"utf8"> means UTF-8 in its former sense, which was liberal and loose and
1019 lax. C<Encode> version 2.10 or later thus groks this subtle but critically
1020 important distinction between C<"UTF-8"> and C<"utf8">.
1022 encode("utf8", "\x{FFFF_FFFF}", 1); # okay
1023 encode("UTF-8", "\x{FFFF_FFFF}", 1); # croaks
1025 In the C<Encode> module, C<"UTF-8"> is actually a canonical name for
1026 C<"utf-8-strict">. That hyphen between the C<"UTF"> and the C<"8"> is
1027 critical; without it, C<Encode> goes "liberal" and (perhaps overly-)permissive:
1029 find_encoding("UTF-8")->name # is 'utf-8-strict'
1030 find_encoding("utf-8")->name # ditto. names are case insensitive
1031 find_encoding("utf_8")->name # ditto. "_" are treated as "-"
1032 find_encoding("UTF8")->name # is 'utf8'.
1034 Perl's internal UTF8 flag is called "UTF8", without a hyphen. It indicates
1035 whether a string is internally encoded as "utf8", also without a hyphen.
1039 L<Encode::Encoding>,
1040 L<Encode::Supported>,
1045 L<perlunicode>, L<perluniintro>, L<perlunifaq>, L<perlunitut>
1047 the Perl Unicode Mailing List L<http://lists.perl.org/list/perl-unicode.html>
1051 This project was originated by the late Nick Ing-Simmons and later
1052 maintained by Dan Kogai I<< <dankogai@cpan.org> >>. See AUTHORS
1053 for a full list of people involved. For any questions, send mail to
1054 I<< <perl-unicode@perl.org> >> so that we can all share.
1056 While Dan Kogai retains the copyright as a maintainer, credit
1057 should go to all those involved. See AUTHORS for a list of those
1058 who submitted code to the project.
1062 Copyright 2002-2014 Dan Kogai I<< <dankogai@cpan.org> >>.
1064 This library is free software; you can redistribute it and/or modify
1065 it under the same terms as Perl itself.