Language code to language name mapping

[Behdad Esfahbod <behdad behdad org>]

Hi everyone,

Hope everyone is having a good time, many of you preparing to be stuffed over
the weekend while the rest of us hack hard :P.

To the issue: Pango currently uses ISO639 language tags to represent
languages.  That has served Pango really well and there's no problem there.
However, many applications need to get to a user-readable representation of
the language code.

I know the iso-codes package provides translations for language and country
names among other things, but I have failed to find the canonical way to go
from a language/country code to language/country name first.  Ideally I think
iso-codes package should abuse the gettext infrastructure to do, such that, say:

  dgettext("iso_639", "fr")

works as

  dgettext("iso_639", "French")

does.  Or maybe add new domains for such lookup.

My question right now is, how do applications (intlclock, etc...) do that
currently?  And how do people think it should be done.

For languages I don't really mind adding the mapping to Pango, but I'd rather
keep it out and in the same place as country mappings.

I have CC'ed the iso-codes maintainers to get their input on this.  Be nice to
them ;).


Cheers,

behdad