Re: Multibyte improvement - g_unichar_to_utf8()

[Yukihiro Nakai <ynakai redhat com>]

> I'd rather not do this in gnumeric.  We use nl_langinfo with
> different argument types than this stub implementation supports.
> Our current approach is to use a configure test for the existence of
> nl_langinfo and to ifdef a fall back to a stupid behaviour when it is
> not supported.
>
> > So there are some policies to development GNOME apps:
> >
> > 1. Complete I18N but ignore *BSD
> > 2. Development also for *BSD users, but just for single byte only.
> > 3. Both
> > 4. Neither
>
> Please explain these further, I'm not clear on what you use the
> decide the different categories.

I means:
If you implement true I18N with useful standard functions in glibc,
it won't be built on *BSD. If you make sure it works on *BSD,
it can't use I18N functions.

But I see you're saying you'll use HAVE_LANGINFO_H macro and try the
more challenging 3...

---
Yukihiro Nakai, Red Hat Japan, Development.