The problem here is visible on application net-misc/mirall. It names the czech translations as cs_CZ, but only valid LINGUA variable is cs. This leads to (by-default) not installing czech translations for me. I suppose there should be some way to allow setting mapping of the translations to allow us to rename the crazily set upstream translations to make them work. The same apply also for sk_SK which is just sk, etc etc...
(In reply to comment #0) > The problem here is visible on application net-misc/mirall. > > It names the czech translations as cs_CZ, but only valid LINGUA variable is > cs. Just FYI, similar problem(about russian local) was discussed in bug #429384
By default all linguas should be used in the short ISO (sk, ru, etc), and only if there are more variants en_GB en_US it should be expanded with default over the basic one so there would be en-> en_US; en_GB -> en_GB.
(In reply to comment #2) > By default all linguas should be used in the short ISO (sk, ru, etc), and > only if there are more variants en_GB en_US it should be expanded with > default over the basic one so there would be en-> en_US; en_GB -> en_GB. Care to elaborate? Why should the short forms be used, and how should they be expanded/mapped to long forms? I have chosen to keep the eclass simple, and just follow upstream naming choices. But patches can be accepted.
(In reply to comment #3) > (In reply to comment #2) > > By default all linguas should be used in the short ISO (sk, ru, etc), and > > only if there are more variants en_GB en_US it should be expanded with > > default over the basic one so there would be en-> en_US; en_GB -> en_GB. > > Care to elaborate? Why should the short forms be used, and how should they > be expanded/mapped to long forms? > > I have chosen to keep the eclass simple, and just follow upstream naming > choices. But patches can be accepted. Census is that language uses ISO 639-1 code unless there is need for more granluation where the default mutation still keeps that ISO and others use the long syntax or stick to the other mutations of iso639.
Dropping kde from cc as the mentioned package is not maintained by us anymore and we don't depend on the l10n.eclass.
Since no patches or proposals on how to tackle this issue have been provided, I am closing this as "won't fix" for now. Please feel free to reopen when there is a proposal or patch.