2 our $VERSION = do { my @r = (q$Revision: 1.37 $ =~ /\d+/g); sprintf "%d."."%02d" x $#r, @r };
10 Carp::croak("encoding pragma does not support EBCDIC platforms");
15 eval { require PerlIO::encoding };
17 $HAS_PERLIO = (PerlIO::encoding->VERSION >= 0.02);
24 $name ||= $ENV{PERL_ENCODING};
26 my $enc = find_encoding($name);
27 unless (defined $enc) {
29 Carp::croak("Unknown encoding '$name'");
31 unless ($arg{Filter}){
32 ${^ENCODING} = $enc; # this is all you need, actually.
33 $HAS_PERLIO or return 1;
34 for my $h (qw(STDIN STDOUT)){
36 unless (defined find_encoding($arg{$h})) {
38 Carp::croak("Unknown encoding for $h, '$arg{$h}'");
40 eval { binmode($h, ":encoding($arg{$h})") };
42 unless (exists $arg{$h}){
44 no warnings 'uninitialized';
45 binmode($h, ":encoding($name)");
55 defined(${^ENCODING}) and undef ${^ENCODING};
57 require Filter::Util::Call ;
58 Filter::Util::Call->import ;
63 if (($status = filter_read()) > 0){
64 $_ = $enc->decode($_, 1);
70 # warn "Filter installed";
72 return 1; # I doubt if we need it, though
79 binmode(STDIN, ":raw");
80 binmode(STDOUT, ":raw");
85 if ($INC{"Filter/Util/Call.pm"}){
86 eval { filter_del() };
97 encoding - allows you to write your script in non-ascii or non-utf8
101 use encoding "greek"; # Perl like Greek to you?
102 use encoding "euc-jp"; # Jperl!
104 # or you can even do this if your shell supports your native encoding
106 perl -Mencoding=latin2 -e '...' # Feeling centrally European?
107 perl -Mencoding=euc-kr -e '...' # Or Korean?
111 # A simple euc-cn => utf-8 converter
112 use encoding "euc-cn", STDOUT => "utf8"; while(<>){print};
114 # "no encoding;" supported (but not scoped!)
117 # an alternate way, Filter
118 use encoding "euc-jp", Filter=>1;
120 # now you can use kanji identifiers -- in euc-jp!
124 Let's start with a bit of history: Perl 5.6.0 introduced Unicode
125 support. You could apply C<substr()> and regexes even to complex CJK
126 characters -- so long as the script was written in UTF-8. But back
127 then, text editors that supported UTF-8 were still rare and many users
128 instead chose to write scripts in legacy encodings, giving up a whole
129 new feature of Perl 5.6.
131 Rewind to the future: starting from perl 5.8.0 with the B<encoding>
132 pragma, you can write your script in any encoding you like (so long
133 as the C<Encode> module supports it) and still enjoy Unicode support.
134 You can write code in EUC-JP as follows:
136 my $Rakuda = "\xF1\xD1\xF1\xCC"; # Camel in Kanji
137 #<-char-><-char-> # 4 octets
138 s/\bCamel\b/$Rakuda/;
140 And with C<use encoding "euc-jp"> in effect, it is the same thing as
143 my $Rakuda = "\x{99F1}\x{99DD}"; # two Unicode Characters
144 s/\bCamel\b/$Rakuda/;
146 The B<encoding> pragma also modifies the filehandle disciplines of
147 STDIN and STDOUT to the specified encoding. Therefore,
149 use encoding "euc-jp";
150 my $message = "Camel is the symbol of perl.\n";
151 my $Rakuda = "\xF1\xD1\xF1\xCC"; # Camel in Kanji
152 $message =~ s/\bCamel\b/$Rakuda/;
155 Will print "\xF1\xD1\xF1\xCC is the symbol of perl.\n",
156 not "\x{99F1}\x{99DD} is the symbol of perl.\n".
158 You can override this by giving extra arguments; see below.
164 =item use encoding [I<ENCNAME>] ;
166 Sets the script encoding to I<ENCNAME>. Filehandle disciplines of
167 STDIN and STDOUT are set to ":encoding(I<ENCNAME>)". Note that STDERR
170 If no encoding is specified, the environment variable L<PERL_ENCODING>
171 is consulted. If no encoding can be found, the error C<Unknown encoding
172 'I<ENCNAME>'> will be thrown.
174 Note that non-STD file handles remain unaffected. Use C<use open> or
175 C<binmode> to change disciplines of those.
177 =item use encoding I<ENCNAME> [ STDIN =E<gt> I<ENCNAME_IN> ...] ;
179 You can also individually set encodings of STDIN and STDOUT via the
180 C<< STDIN => I<ENCNAME> >> form. In this case, you cannot omit the
181 first I<ENCNAME>. C<< STDIN => undef >> turns the IO transcoding
186 Unsets the script encoding. The disciplines of STDIN, STDOUT are
187 reset to ":raw" (the default unprocessed raw stream of bytes).
195 The pragma is a per script, not a per block lexical. Only the last
196 C<use encoding> or C<no encoding> matters, and it affects
197 B<the whole script>. However, the <no encoding> pragma is supported and
198 B<use encoding> can appear as many times as you want in a given script.
199 The multiple use of this pragma is discouraged.
201 Because of this nature, the use of this pragma inside the module is
202 strongly discouraged (because the influence of this pragma lasts not
203 only for the module but the script that uses). But if you have to,
204 make sure you say C<no encoding> at the end of the module so you
205 contain the influence of the pragma within the module.
207 =head2 DO NOT MIX MULTIPLE ENCODINGS
209 Notice that only literals (string or regular expression) having only
210 legacy code points are affected: if you mix data like this
214 the data is assumed to be in (Latin 1 and) Unicode, not in your native
215 encoding. In other words, this will match in "greek":
221 "\xDF\x{100}" =~ /\x{3af}\x{100}/
223 since the C<\xDF> (ISO 8859-7 GREEK SMALL LETTER IOTA WITH TONOS) on
224 the left will B<not> be upgraded to C<\x{3af}> (Unicode GREEK SMALL
225 LETTER IOTA WITH TONOS) because of the C<\x{100}> on the left. You
226 should not be mixing your legacy data and Unicode in the same string.
228 This pragma also affects encoding of the 0x80..0xFF code point range:
229 normally characters in that range are left as eight-bit bytes (unless
230 they are combined with characters with code points 0x100 or larger,
231 in which case all characters need to become UTF-8 encoded), but if
232 the C<encoding> pragma is present, even the 0x80..0xFF range always
235 After all, the best thing about this pragma is that you don't have to
236 resort to \x{....} just to spell your name in a native encoding.
237 So feel free to put your strings in your encoding in quotes and
240 =head2 tr/// with ranges remain unaffected
242 The B<encoding> pragma works by decoding string literals in
243 C<q//,qq//,qr//,qw///, qx//> and so forth. As of perl 5.8.0, this
244 does not apply to C<tr///>. Therefore,
246 use encoding 'euc-jp';
248 $kana =~ tr/\xA4\xA1-\xA4\xF3/\xA5\xA1-\xA5\xF3/;
249 # -------- -------- -------- --------
253 $kana =~ tr/\x{3041}-\x{3093}/\x{30a1}-\x{30f3}/;
257 =item Legend of characters above
259 utf8 euc-jp charnames::viacode()
260 -----------------------------------------
261 \x{3041} \xA4\xA1 HIRAGANA LETTER SMALL A
262 \x{3093} \xA4\xF3 HIRAGANA LETTER N
263 \x{30a1} \xA5\xA1 KATAKANA LETTER SMALL A
264 \x{30f3} \xA5\xF3 KATAKANA LETTER N
268 =head3 workaround to tr///;
270 You can, however, achieve the same as simply as follows;
272 use encoding 'euc-jp';
274 eval qq{ \$kana =~ tr/\xA4\xA1-\xA4\xF3/\xA5\xA1-\xA5\xF3/ };
276 Note the C<tr//> expression is surronded by C<qq{}>. The idea behind
277 is the same as classic idiom that makes C<tr///> 'interpolate'.
279 tr/$from/$to/; # wrong!
280 eval qq{ tr/$from/$to/ }; # workaround.
282 Nevertheless, in case of B<encoding> pragma even C<q//> is affected so
283 C<tr///> not being decoded was obviously against the will of Perl5
284 Porters. In future version of perl, this counter-intuitive behaviour
285 of C<tr///> will be fixed so C<eval qq{}> trick will be unneccesary.
287 =head1 Non-ASCII Identifiers and Filter option
289 The magic of C<use encoding> is not applied to the names of
290 identifiers. In order to make C<${"\x{4eba}"}++> ($human++, where human
291 is a single Han ideograph) work, you still need to write your script
292 in UTF-8 or use a source filter.
294 In other words, the same restriction as with Jperl applies.
296 If you dare to experiment, however, you can try the Filter option.
300 =item use encoding I<ENCNAME> Filter=E<gt>1;
302 This turns the encoding pragma into a source filter. While the default
303 approach just decodes interpolated literals (in qq() and qr()), this
304 will apply a source filter to the entire source code. In this case,
305 STDIN and STDOUT remain untouched.
309 What does this mean? Your source code behaves as if it is written in
310 UTF-8. So even if your editor only supports Shift_JIS, for example,
311 you can still try examples in Chapter 15 of C<Programming Perl, 3rd
312 Ed.>. For instance, you can use UTF-8 identifiers.
314 This option is significantly slower and (as of this writing) non-ASCII
315 identifiers are not very stable WITHOUT this option and with the
316 source code written in UTF-8.
318 To make your script in legacy encoding work with minimum effort,
319 do not use Filter=E<gt>1.
321 =head1 EXAMPLE - Greekperl
323 use encoding "iso 8859-7";
325 # \xDF in ISO 8859-7 (Greek) is \x{3af} in Unicode.
330 printf "%#x\n", ord($a); # will print 0x3af, not 0xdf
334 # $c will be "\x{3af}\x{100}", not "\x{df}\x{100}".
336 # chr() is affected, and ...
338 print "mega\n" if ord(chr(0xdf)) == 0x3af;
340 # ... ord() is affected by the encoding pragma ...
342 print "tera\n" if ord(pack("C", 0xdf)) == 0x3af;
344 # ... as are eq and cmp ...
346 print "peta\n" if "\x{3af}" eq pack("C", 0xdf);
347 print "exa\n" if "\x{3af}" cmp pack("C", 0xdf) == 0;
349 # ... but pack/unpack C are not affected, in case you still
350 # want to go back to your native encoding
352 print "zetta\n" if unpack("C", (pack("C", 0xdf))) == 0xdf;
354 =head1 KNOWN PROBLEMS
356 For native multibyte encodings (either fixed or variable length),
357 the current implementation of the regular expressions may introduce
358 recoding errors for regular expression literals longer than 127 bytes.
360 The encoding pragma is not supported on EBCDIC platforms.
361 (Porters who are willing and able to remove this limitation are
366 L<perlunicode>, L<Encode>, L<open>, L<Filter::Util::Call>,
368 Ch. 15 of C<Programming Perl (3rd Edition)>
369 by Larry Wall, Tom Christiansen, Jon Orwant;
370 O'Reilly & Associates; ISBN 0-596-00027-8