Re: Crashes with noveau driver and non-Debian software: How can I report the problems?
Hi, On 18.08.2017 14:22, Greg Wooledge wrote: > Any package beginning with task- is a meta-package which can be safely > removed. The sole purpose of a meta-package is to bring in other > packages as dependencies. It has no content of its own. is removal also safe in case of system upgrades? For example, could the upgrade process neglect to pull in new packages because the meta-package is missing? Kind regards, Erik Brangs
Re: Crashes with noveau driver and non-Debian software: How can I report the problems?
Hi, On 16.08.2017 13:08, Sven Joachim wrote: > No, purging xserver-xorg-video-nouveau won't have the desired effect, > because the modesetting driver will also load nouveau_dri.so. That's good to know. I don't want to purge xserver-xorg-video-nouveau because Synaptic tells me that this would require removing task-desktop, task-xfce-desktop and xserver-xorg-video-all. The names of those packages sound like they should be kept. Should I report a bug against xserver-xorg-video-nouveau to report the crashes? Or is that not appropriate because the affected binary isn't shipped by Debian? Kind regards, Erik Brangs
Crashes with noveau driver and non-Debian software: How can I report the problems?
Hi, I'm using Anki as a flashcard software. It was available in Jessie but has been removed for Stretch because upstream didn't release a version for current QT in time for Stretch. It's unlikely that Anki can get backported to Stretch because that would require a backport of QT [1]. So I need to get Anki from a non-Debian source. There are binary builds available at the upstream site at https://apps.ankiweb.net/docs/beta.html and they run on Stretch machines. The binaries sometimes crash on my machine and the console shows error messages from the nouveau driver. I asked upstream about this and they said it was "likely a buggy GL implementation" in the graphics driver. The binaries work fine if I disable hardware acceleration using LIBGL_ALWAYS_SOFTWARE=1 . I'd like to know if I should report any bugs about this and/or which packages I could install from backports to check if the driver problem is gone there. Kind regards, Erik Brangs [1] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=855942