Naming new locales (was Re: Improving the GNU libc locales)

[Danilo Segan <dsegan at gmx dot net>]

Petter Reinholdtsen <pere@hungry.com> writes:

> [Danilo Segan]
>> Does that mean that I should submit my sr_CS* locales (I've got a new
>> one, for Serbian Jekavian dialect -- all available from
>> http://srpski.org/locale/) to Bugzilla system now?
>
> Yes.

Ok, I'll submit them right away.

This is another question that should probably be asked here.

The new locale, "Serbian Jekavian" is still pretty much unused (just 8
module translations have been so far submitted), so I don't have any
firm position on the naming issue.

Yet, I'd like it to be included in GNU libc as well.  Because of all
the troubles experienced with sr_CS@Latn, I don't want to have naming
issues delay this, so I'll ask in front:

  What modifier name should I use for this locale?

It will be in Cyrillic script (perhaps there'll be a need for Latin
in the future, but I don't plan in the short run), and Serbian name
for dialect is "Ijekavski".  That's why Bojan Suzic, Jekavian speaker
who's doing the translations, suggested "sr@ije".

Would there be anything against this, and should I choose something
else (the earlier this is decided, the better).  I can imagine Ulrich
preferring "sr_CS@jekavian", but I have so far respected translators'
decision.

The good thing about GNU libc's setlocale() implementation is that it
will use basic locale if no @modifier is found (that's in
libc/locale/findlocale.c).  Still, naming of days is different in
this locale, so it needs to be installed separately as well.

Cheers,
Danilo