Re: How to do the language code switch for Norwegian Bokmål: no -> nb

[Kjartan Maraas <kmaraas broadpark no>]

man, 02,.08.2004 kl. 19.21 +0200, skrev Christian Rose:
> mån 2004-08-02 klockan 17.20 skrev Åsmund Skjæveland:
> > Some within the Norwegian translation teams are now arguing that it's 
> > time for GNOME to use 'nb' as the language code for Norwegian Bokmål 
> > rather than the obsolete 'no'. KDE has long since switched, and libc now 
> > includes 'nb' as a real locale as well as 'no'. Kjartan Maraas is AFAIK 
> > in charge, but he hasn't been heard from in a while.

Where has this been argued? I seem to have been unsubscribed from the
l10n lists on lister.ping.uio.no if that's the place...

> > 	What is the best way to do this, technically? Some argue that the 
> > install scripts of the GNOME apps should be modified so that 'nb.po' is 
> > installed both as 'no.mo' and 'nb.mo', so that there is only one po 
> > file, but two working locale codes in a transition period.
> > 	If this approach requires the cooperation of all maintainers, it's 
> > probably not the best way to go, but if it can be included in intltool 
> > and make sure all GNOME releases then depend on a recent intltool, it 
> > could work.
> > 	Having two po files 'no.po' and 'nb.po' and letting the translators 
> > make sure they're identical is apparently not a desirable option. 
> > Probably the sentiment is that then the translators have to do this 
> > extra step for every commit, there's always a chance that someone will 
> > forget, while changin the build script to copy nb.po to no.po would only 
> > have to be done once.
> > 
> > Can it be done for GNOME 2.8?
> 
I'm tempted to wait for the next release cycle and then rename from
no.po to nb.po. This will let us have ample time to test stuff along the
way. I think it would be easiest if we make scripts to rename stuff in
CVS directly and then change configure.[in|ac](ALL_LINGUAS) to reflect
the renaming.

> If it involves some hacks to make the transition work smoothlessly, I'd
> say it's something that should be made early in the development cycle
> rather than very late, so that it can be properly tested. So, given
> this, perhaps testing this in HEAD as soon as the modules branch for
> GNOME 2.8 would be better than including this in GNOME 2.8 at this
> point. But I'm not sure I understand every aspect of what needs to be
> involved here, so that's just my uninformed opinion.
> 
I tend to agree with this point of view.

> > How much work, roughly, would be involved to make it work?
> 
> Kjartan, can you comment?
> 
I think we should make some sort of assessment of how many of the major
distros have upgraded to the libc that contains the new locale. It would
help to get some data on how this would affect Solaris and possibly
other operating systems too just to be able to inform people who end up
experiencing problems.

Cheers
Kjartan