Re: Accepted po-debconf 0.2.2 (all source)
On Mon, 16 Sep 2002, Junichi Uekawa wrote:
> > > Not really, there is only one charset used in templates files (EUC-JP),
> > > but user may have a different charset. This is not only true for Japanese,
> > > but all other languages may be in UTF-8. What we need is charset conversion
> > > at runtime.
> >
> > It would be somewhat elegant if debconf and the eventual translated
> > package descriptions end up using the same scheme for storing the
> > translations, which I thought was informally agreed on to be universally
> > UTF-8, and converted at runtime for display.
>
> I have had an impression that we didn't need to universally make everything
> UTF-8. We can convert from any charset to any other charset (almost).
The reason people want to use UTF-8 for package descriptions is so that
we don't have the Evil multiple charsets in a single file problem. The
control file, Package File, etc, etc, call all be purely in UTF-8 and
store as many different languages as they want. This is a desired
feature.
If that is how the package file is going to be handled then I don't see
any reason that debconf should deviate and support various encodings.
You don't have to throw away any existing non-UTF8 translations since they
can easially be converted.
Jason
Reply to: