Languagestats-stats-for
No, it seems that $1 is the feedback language because the German language name appears for Upper Sorbian and Lower Sorbian statistics.
Not sure what you mean. Please provide full URL, including uselang parameter, and then explain what you think is wrong.
Well, the text of message Languagestats-stats-for is Translation statistics for $1 ($2). It appears in Language statistics above the table, see Upper Sorbian stats. The Upper Sorbian text of this message is: Přełožowanska statistika za $1 ($2). For parameter $1 the Upper Sorbian language name should appear. But the German language name appears instead, thus Přełožowanska statistika za Obersorbisch (nowše přełožki). It should be Přełožowanska statistika za hornjoserbšćinu (nowše přełožki).
I thought Der Umherirrende meant this issue. He wrote that it is always the English language name. But for Upper Sorbian and Lower Sorbian the language name appears in German. Therefore I suppose that feedback language is used.
That's because the Uppersorbian language translations are not in CLDR 1.8, so it takes the translation from the first available fallback in CLDR, being German.
Thanks. CLDR is used in MediaWiki? I always thought that the file names.php is used for defining localized language names. I know that in any external translation project of translatewiki CLDR plural rules are used, but inside MediaWiki?
Names.php only contains autonyms, no translations. The MediaWiki extension CLDR has the available translations from CLDR for language names.
Thanks. What a pity. Maybe I will contact CLDR and do all what is necessary for registering Upper Sorbian and Lower Sorbian. I've already seen xml files that contain the language names in the respective language. They are app. 3650 lines big. It will be a lot of work and this for 2 languages.