| Subject: | Wrong usage of encode_utf8 |
Using encode_utf8 in the PRINT method of Apache::OutputChain is wrong.
The comment behind claims that it's a no-op but it isn't, as high-bit
characters would turn into utf-8 encoded characters. It's better to
just check if the utf8 flag is set in the string and just remove the flag.
Maybe it would also be appropriate to issue a warning a la "wide character
in PRINT", because I consider it an error to have utf8-flagged characters
in this stage.
Regards,
Slaven
| Subject: | Apache-GzipChain.diff |
--- /usr/local/lib/perl5/site_perl/5.8.8/Apache/GzipChain.pm Mon Mar 3 12:49:42 2003
+++ /home/e/eserte/trash/Apache/GzipChain.pm Wed Aug 8 17:45:20 2007
@@ -2,7 +2,7 @@ package Apache::GzipChain;
use Compress::Zlib 1.0;
use strict;
use vars qw(@ISA $VERSION);
-$VERSION = sprintf "%d.%02d", q$Revision: 1.14 $ =~ /(\d+).(\d+)/;
+$VERSION = sprintf "%d.%02d", q$Revision: 1.15 $ =~ /(\d+).(\d+)/;
use Apache::OutputChain;
@ISA = qw( Apache::OutputChain );
@@ -35,7 +35,9 @@ sub PRINT {
return unless length($res);
if ($] > 5.007) {
require Encode;
- $res = Encode::encode_utf8($res); # noop if $res doesn't have the UTF-8 flag set
+ if (Encode::is_utf8($res)) {
+ Encode::_utf8_off($res);
+ }
}
$self->Apache::OutputChain::PRINT(Compress::Zlib::memGzip($res));
}