When you set the Time to use C (default) the locale is set to LC_TIME=C.UTF-8. Since there is no C.UTF-8 locale some commands generate warning messages like "Failed to set locale, defaulting to C". Reproducible: Always Steps to Reproduce: 1. Set Time to use C in systemsettings 2. Type locale and see the bad settings 3. Actual Results: LC_TIME=C.UTF-8 Expected Results: LC_TIME=C
Here https://bugzilla.redhat.com/show_bug.cgi?id=1250238 patch to fix this.
I took a quick look and don't see where in the Date & Time kcm to set the Time to use C? where did you do that?
Created attachment 94507 [details] C locale in settings Look at the screenshot.
From that bug report at redhat it seems the issue is a redhat/fedora specific issue. I'm able to use the Time "Default (C)" here and see no errors about having C.UTF-8 in my locale. It seems to be because redhat/fedora doesn't have that locale yet. It appears the customization has happened in a fedora package patch, so shouldn't be a problem anymore.
From my understanding, the characterization in comment #4 is incorrect. My understanding is: * yes, fedora/redhat doesn't have that locale yet, but that is because... (stock/upstream) glibc doesn't support that locale yet either, see https://sourceware.org/bugzilla/show_bug.cgi?id=17318 * Some other distributions are patching in support for it themselves, for example, http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=636086 Handling both cases understandably could make *this* bug more difficult. I would argue kde should prioritize support for the stock case (of course, I'm biased).
I see, if KDE adopts the patch from upstream will that cause any trouble with distributions that do have C.UTF-8 locale support? If not I say yeah, we should add it (and I will, np)
s/upstream/redhat bug report/
The impact of adopting this change will be that LANG (and friends) will end up being "C" instead of "C.UTF-8". I'm not a locale expert, and it *is* a change of behavior which could be argued to qualify as "trouble", but not having this change is also the cause of trouble for unpatched distros too.
Fwiw, looks like fedora finally adopted C.UTF-8 in it's rawhide development branch (uncertain about gliibc upstream status), so as far as *I* am concerned, we can close this without action now.