Allow all locale IDs we use
Language code can have three characters (e.g. ast). We also use the script (e.g. for Serbian). And for Interlingua we need the 3-digit region code. We don't and probably never will use the extensions (-u-nu-*) that Android supports. Therefore, I left them out.
While browsing the logs of fdroidpublisher I noticed that we filter out the Asturian localization. Moreover, Interlingua is now supported by Android and we do use the script IDs for Serbian. The extended regular expression should match all locale IDs that are used by us.