On Sun, Feb 04, 2007, Joey Hess wrote:
> > I could not find a way to get the encoding of a locale reliably, so
> > perhaps you can use is_utf8 && decode_utf8 for all strings before
> > outputting them?
> The entire point of use open qw{:utf8 :std} is to avoid having to
> manually encode every single string that the program outputs.
This was a suggested workaround...
> This seems to be basically a bug in perl, that it provides no way to use
> utf8 for all IO if you also want to use $!, $@ or other variables of
> this sort, since its utf8 support does not include converting those
> values to utf8.
Yes, I agree it's likely a bug in Perl, however I don't know whether
Perl can reliably tell the encoding of $!, so it might be a limitation
that if you want to output localized $!, you need to avoid setting the
encoding of the output file descriptor.
--
Loïc Minier <[EMAIL PROTECTED]>
