On Sun, Mar 18, 2012 at 6:03 AM, Christian PERRIER <bubu...@debian.org>wrote:
> > I'm afraid, this is an incorrect use of locales. > > Country modifiers shouldn't be used for language variants. So a ug_US > locale means nothing (Uyghur in United States of America? Why not > ug_UK or ug_PT or whatever?). > > The correct locale for "Uyghur written in latin script" should be > "ug_CN@latin" (and eventually ug_XX@latin, in case Uyghur is widely > used in another country than "China", such as Kazakhstan, for > instance. > > And, in case Uyghur is also written in Cyrillic, as you seem to imply, > then the right locale would be ug_XX@cyrillic, where XX is the country > where the cyrillic variant is the most widely used .....I guess KZ, > then > > I would strongly advise against using ug_KZ to denote "Uyghur written > in cyrillic ". It should be kept for "Uyghur written in Arabic script, > in Kazakhstan" (and if that means nothing as Uyghur is never written > in arabic script in KZ, then don't create the locale but create > ug_KZ@cyrillic). > > But, really, ug_US should not happen and I doubt upstream glibc > maintainers accept it. As long as they use different country code, it will work. ug_CN ug_KZ@cyrillic ug_US@latin (or ug_XX@latin, XX other than CN or KZ) For the Latin based Uyghur, United States is the country where it is used heavily.