Please describe the problem:
When building its locale list for the language selector, it seems that gdm does not infer Serbian locales properly. This is on Fedora 10, gdm 2.24.0.

The total number of infered locales is correct, but there are some deficiencies as shown in the mapping below:

1. српски (Србија) -> sr_RS.UTF-8@latin (locale is correct, but label should appear in Latin script as "srpski (Srbija)")

2. српски (Србија) -> sr_RS.latin (invalid locale, should be sr_RS.UTF-8)

3. српски (Србија и Црна Гора) -> sr_ME.UTF-8 (locale is correct, but label should be "српски (Црна Гора)")

Problems 1 and 3 are I guess due to missing and outdated iso-codes translations, and can be easily fixed upstream by our community, and I have already verified a fix by building custom iso-codes translations.

That leaves 2 as the real bug. Users cannot get to sr_RS.UTF-8 in any combination above.

Available Serbian locales in Fedora 10 are

$ locale -a|grep sr
sr_ME
sr_ME.utf8
sr_RS
sr_RS@latin
sr_RS.utf8
sr_RS.utf8@latin

Steps to reproduce:



Actual results:


Expected results:


Does this happen every time?


Other information: