[07:06] <tjaalton> RAOF: hi, still around?
[07:33] <duflu> tjaalton: I think he's on vacation for a while. Maybe back at some point this week briefly
[07:35] <tjaalton> duflu: ok, there was just a thing with the mir egl patch for mesa, missing 'break;'s since an update to it 3y ago, wondering if it was a mistake
[07:36] <tjaalton> uploaded 12.0.1 btw
[07:37] <duflu> tjaalton: I have a collection of bugs in the mir egl patch :)  https://bugs.launchpad.net/ubuntu/+source/mesa/+bugs?field.tag=egl-platform-mir
[07:37] <duflu> Woo Mesa 12
[07:38] <tjaalton> needed just a slight rebase, and noticed these
[07:40] <tjaalton> duflu: oh, well I can fix 1473091 on the distro
[07:41] <duflu> tjaalton: You mean 1473901? You're my hero
[07:41] <tjaalton> ah right
[07:42] <tjaalton> but doesn't dropping those affect other users of gm_dri_is_format_supported?
[07:42] <tjaalton> gbm_*
[07:43] <duflu> tjaalton: It's just a mistake we introduced in the patch. I think it's safe for us to remove it
[07:43] <duflu> Certainly it would be nice for it to finally work
[07:43] <tjaalton> oh right
[07:47] <tjaalton> actually, upstream added XBGR8888 there, so our patch no longer carries that
[07:48] <tjaalton> that's new with 12
[07:49] <duflu> tjaalton: Added it in Mesa 12.x?
[07:49] <duflu> As in it should work properly now?
[07:50] <duflu> I know a certain Xmir bug that might benefit from that
[07:50] <tjaalton> yes
[07:51] <duflu> Oooh, I thought it would never happen
[07:53] <duflu> tjaalton: Great news assuming it's not a mistake. XBGR8888 is the same thing as RGBX for Android. But why would Mesa support XBGR but not ABGR?
[07:56] <tjaalton> well, ABGR was already added but not to that function so not really useable
[07:56] <tjaalton> aiui
[08:12] <duflu> tjaalton: Still one of those switch cases is wrong
[08:12] <duflu> Which is better than two
[08:12] <tjaalton> ok I'll drop the other one
[08:13] <duflu> tjaalton: Yeah at least till upstream adds it intentionally.
[08:13] <duflu> It might work now, but that would only be by accident
[08:13] <tjaalton> and then we'll get it for free
[08:15] <duflu> tjaalton: Looks like a possible upstream bug in the April 2016 change though. They should have added GBM_FORMAT_ABGR8888 at the same time
[08:15] <duflu> Because Android requires it
[08:18] <tjaalton> ah
[08:20] <duflu> But simply having a case statement for it may not be enough. Safer to omit it till it's fully implemented upstream
[08:21] <duflu> Could be enough for Mir. Not enough for other projects
[08:21] <tjaalton> right
[08:53] <alan_g> anpok: did you have a chance to try https://code.launchpad.net/~alan-griffiths/miral/fix-1603086/+merge/300220?
[08:53] <robert_ancell> duflu, what did you have in mind with rethinking how LightDM launches USC?
[08:55] <duflu> robert_ancell: No idea :) It's a requirement rather than a design
[08:55] <robert_ancell> duflu, what's the requirement?
[08:56] <duflu> robert_ancell: Just easier control/restarting of unity-system-control independently of lightdm
[08:56] <robert_ancell> duflu, that
[08:57] <robert_ancell> that's a pretty severe operation - do you expect the sessions to survive?
[08:57] <duflu> robert_ancell: OK, nevermind. I found unity-compositor-command=/usr/bin/env ... works
[08:57] <robert_ancell> nice
[08:58] <duflu> robert_ancell: Hmm, though could you simply stop lightdm from deleting the environment of unity-system-compositor?
[08:58] <robert_ancell> duflu, you mean inherit the lightdm environment?
[08:58] <duflu> robert_ancell: Yeah I guess
[09:00] <robert_ancell> it's worth considering
[09:00] <alan_g> duflu: you should also be able to create unity-system-compositor.conf to put options in.
[09:00] <duflu> alan_g: I was wondering that but found no default file existed
[09:01] <alan_g> I guess no-one has found a need for it yet
[09:02] <duflu> I'd still rather not publicise file formats at all till there's a strong need to standardize them
[09:03] <duflu> Oh. I could also have given --options to unity-compositor-command
[09:04] <anpok> alan_g: oh it seems I have been using it all day now
[09:06] <alan_g> So the latest version (from Friday) works for your use case?
[09:07] <anpok> yes
[09:43] <duflu> alan_g, anyone, remember the workaround for "GtkSettings Cursor Theme: Unsupported GDK backend
[09:43] <duflu> " ?
[09:44] <alan_g> how are you getting that?
[09:45] <duflu> alan_g: nautilus via gtk on Mir (Unity8)
[09:47] <duflu> Oh that's not the crash. Just a message that precedes the crash
[09:47] <duflu> Nevermind
[09:47] <alan_g> Mmm... may not be the thing you've hit. But there are issues running gkt apps as root (that I've yet to dig into). I use -arw-file and run as a normal user.
[09:50] <duflu> Not running as root.
[09:51] <duflu> anpok: Hey can you add 'nautilus' to your list of GTK apps to test? :)
[09:51] <duflu> I just realized I test Xmir with it lots, but never Mir native
[09:55] <duflu> Found it. Turns out that's a bug I fixed elsewhere.
[09:59] <duflu> anpok: If you're interested, I've possibly fixed the same bug elsewhere already: https://bugs.launchpad.net/ubuntu/+source/nautilus/+bug/1603923
[10:33] <anpok> oh yes .. cursor settings is on my list of things to be resolved soonish