On Tuesday 08 December 2009, Christian Perrier wrote: > None of these will trigger the possibility of a double entry in the > language list, so I think that the problematic test can be dropped. OK. Should we maybe add a test that a valid locale "ll_CC.*" exists for the combination, so that we avoid adding really weird entries like en_NL or de_VI, but do add the ones we want like de_LI, nl_BE, en_IE?