[10:20] <tjaalton> pushed mesa 11.2.0-rc1 to ppa:tjaalton/ppa
[10:21] <tjaalton> mir-egl patch didn't need work, and it at least compiles
[10:21] <tjaalton> there is a FFE bug for this, but it will be properly tested before pushing to main
[10:59] <alf_> sil2100: Hi! When you get some free cycles, could you please also disable unity-system-compositor jobs in s-jenkins (unity-system-compositor-ci, unity-system-compositor-autolanding etc)
[11:01] <sil2100> alf_: ok, will try to do that in a minute :)
[11:01] <alf_> sil2100: thanks
[13:49] <Saviq> alf_, can you please add me to your CI team so I can steal your device jobs? :)
[13:49] <alf_> Saviq: sure
[13:51] <Saviq> alf_, oh, maybe I'm there already, just was looking at a config from a multi-config job
[13:51] <alf_> Saviq: so, our 'ci team' is actually mir-team
[13:51] <Saviq> alf_, yeah just saw
[13:51] <alf_> Saviq: so you should be there
[13:52] <Saviq> alf_, looking at device-0-flash, find it weird that you flash per label, shouldn't you have it per slave instead? otherwise you rely on the fact that you have just one of each?
[13:54] <alf_> Saviq: I thought that if I select a label in a multiconfig job it would just pick a single node matching that label?
[13:55] <Saviq> alf_, that's correct, but how do you then know that you flash and test on that same slave?
[13:55] <Saviq> I mean a "monitor" job can have a label, but downstream from that you need to stick to a slave
[13:57] <alf_> Saviq: yeah, you are right
[13:58] <alf_> Saviq: perhaps there is a way to constrain the node to be the same for run job as it is for the flash job?
[13:59] <Saviq> alf_, only way I can think of is for the upstream job to run on that slave (throttled to just one per slave) and run downstream jobs on that same slave
[14:00] <Saviq> alf_, otherwise other jobs could get queued on that slave in the mean time
[14:00] <Saviq> I mean, they still could, if you did it manually
[14:00] <Saviq> other than having a single job that does everything in sequence (flash + test)
[14:01] <Saviq> without letting the slave out
[14:01] <Saviq> I think that's the safer approach actually
[14:02] <Saviq> because otherwise you can queue a job on the same slave before the flashing completes
[14:02] <Saviq> and that would "steal" the slae
[15:16] <alan_g> alf_: if you fix the conflict I'll TA
[15:16] <alan_g> https://code.launchpad.net/~afrantzis/unity-system-compositor/log/+merge/274550
[15:32] <alf_> alan_g: thanks, see latest comment on why it can't land yet
[15:34] <alan_g> /sigh! So many barriers to progress!
[18:58] <fsbif> hello everyone. I can't get unity 8 run in ubuntu 15.10
[18:59] <fsbif> it says "unity8-lxc-setup: error: The container already exists." when I try to remove the container
[19:05] <ChrisTownsend> fsbif: Hi!  How exactly are you trying to remove the unity8-lxc container?
[19:05] <ChrisTownsend> D'oh!