[02:08] <duflu> mterry: We're not tied to my CMake prototype at all. In fact I might context switch now and not get back to this task for another week or so...
[03:37] <RAOF> vgdb-error is basically the best thing ever.
[03:38] <RAOF> Ah, good. It's a race.
[03:38] <RAOF> That'll make it easier.
[03:42] <duflu> Hmm, why is upstart being added *back in* as part of a vivid update?
[03:45] <RAOF> It's still session init, right?
[11:17] <duflu> greyback_: Is there a way to get through the Unity8 welcome wizard with a mouse? I can't seem to do the right edge swipe
[11:17] <duflu> Tried two machines
[11:18] <duflu> Never mind... G'night
[11:18] <greyback_> duflu: hmm, quite possibly a bug.
[11:18] <duflu> Yeah seems like it
[11:18] <duflu> Later...
[11:18] <greyback_> duflu: o/
[11:31] <alf_> Saviq: greyback_: Do any recent unity8/dash updates contain performance improvements?
[11:33] <greyback_> alf_: I can see nothing performance specific in the unity8 log
[11:33] <greyback_> alf_: dash looking good with double-buffering?
[11:34]  * greyback_ would be surprised
[11:34] <greyback_> dash often used to go over 16ms per frame
[11:36] <alf_> greyback_: Yes, I am suprised too, but I will wait for some independent confirmation of my results
[11:37] <greyback_> alf_: ok
[11:37] <greyback_> perhaps some perf improvements landed in qt5 recently
[11:37]  * greyback_ reflashes
[11:39]  * alf_ checks list of updated packages in phone
[11:42] <alf_> greyback_: so there was an upgrade from 5.4.0 to 5.4.1
[11:48] <alf_> greyback_: But I can't see anything interesting in the Qt 5.4.1 changelog
[11:51] <greyback_> alf_: nor do I
[12:08] <greyback_> alf_: think there's 1 fix which has a positive impact on rendering performance, as yeah I see a definite improvement on the devel-proposed image right now
[12:08] <alf_> greyback_: which fix?
[12:11] <greyback_> alf_: https://code.launchpad.net/~aacid/thumbnailer/fix_dbus_blocking/+merge/251065
[12:11] <alf_> greyback_: nice, thanks
[13:41] <kgunn> alf_: trying to sort this one out https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1418456
[13:41] <kgunn> you one comment seems pretty certain it's due to the other 2 bugs you referenced
[13:42] <kgunn> but the 1414999 bug seems fixed in mesa? (so unrelated?)
[13:43] <kgunn> oh nvmd...could be related, i see it's unity8 desktop
[13:49] <alf_> kgunn: right, the core problem is not being able to get an EGL context, possibly because of unsupported gpu (possibly fixed by Mesa 10.5 depending on the gpu). The "double free" part of the problem was an because we didn't tear down the EGL driver correctly when the error happened (fixed).
[13:49] <alf_> s/an because we/that/
[14:05] <anpok> hm I wonder if this patch is still shipped with mesa: http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/vivid/mesa/vivid/view/head:/debian/patches/i915-dont-default-to-2.1.patch
[14:28] <anpok> alan_g: hm what do you mean by adding another level of indirection?
[14:28] <anpok> https://code.launchpad.net/~andreas-pokorny/mir/stub-input-platform/+merge/252760/comments/631055
[14:31] <anpok> i tried to avoid running into the symbol mangling issues caused by having an anonymous type in the signature, and do not see a way to keep the constructor private
[14:31] <alan_g> anpok: well, you could define friend builder function that's called in make_module_ptr, or make the constructor protected and have a local derived class in make_module_ptr
[14:37] <anpok> alan_g: brilliant!
[18:08] <alf_> Saviq: camako: I have trouble running the unity8 autopilot tests on the phone. Has the way we run them changed?
[18:09] <alf_> e.g. phablet-test-run -n -r 0000 -p unity8-autopilot unity8 doesn't work for me (can't find unity8-autopilot)
[18:43] <alf_> greyback_: ^^ Any ideas?
[19:18] <alf_> greyback_: Saviq: camako: arggh, I just had to run apt-get update :/
[19:26] <greyback_> alf_: yeah that'll do it
[19:33] <camako> alf: :-)
[20:48] <racarr> privatizing event header
[20:48] <racarr> feels good :D
[20:49] <racarr> NO MORE INPUT
[20:49] <racarr> I mean privatizing the old event header of course
[20:49] <racarr> ;)