On Mon, 2016-08-08 at 09:16 +0200, Timm Bäder wrote: > That might be possible of course (especially if your code base does not > contain a widget that creates its own GdkWindows), but the advice is the > same for all criticals and warnings you get: Run in gdb with > G_DEBUG=fatal-warnings or G_DEBUG=fatal-criticals and once you hit the > one you're interested in, print a backtrace (make sure you have > GTK+ debugging symbols). So I have to recompile all my GTK libs to get debugging symbols? Well, as you may have seen, I made a minimal plain C example two days ago, which shows a very similar behaviour. https://mail.gnome.org/archives/gtk-list/2016-August/msg00007.html At least now I am sure that it is not related to Nim language. I have some hope that that problem may vanish with GTK 3.20 -- but not too much, because it seems to be never reported to bugzilla. So I think I will wait until 3.20 is available for my Box, and send that example to gtk bugzilla when the problem persists. _______________________________________________ gtk-list mailing list gtk-list@xxxxxxxxx https://mail.gnome.org/mailman/listinfo/gtk-list