[08:58] <anpok_> hum what is the purpose of clients only use client symbols test
[08:58] <anpok_> as in it reports only missing client symbols but still fails
[08:59] <anpok_> ah .. ok
[09:00] <anpok_> gold..
[09:08] <duflu> anpok_: The history behind that was about a year ago we had clients accidentally bound to particular libmircommon symbols. And libmircommon's ABI breaks too often. So test that clients only ever talk to libmirclient
[09:09] <duflu> and it's "more stable" ABI
[09:09] <anpok_> yeah.. it just failed for me after switching to gold
[09:09] <anpok_> and just saw raofs mp
[09:11] <anpok_> which btw eleminates the slow linking problem..
[09:36] <alan_g> anpok_: any thoughts on this? https://code.launchpad.net/~alan-griffiths/mir/fix-1526779/+merge/280880/comments/712083
[09:38] <anpok_> alan_g: whats happening there?
[09:39] <alan_g> Read the bug discussion?
[09:39] <alan_g> s//Have you /
[09:39] <anpok_> not yet
[09:39] <anpok_> oh
[09:43] <alan_g> So, something is screwed up in drm (or below) - and I suspect other kit could be affected.
[09:46] <anpok_> aye..gbm and drm are screwed ..
[09:46] <anpok_> there seems to be no attempt to ever behave as a device abstraction
[10:38] <alan_g> anpok_: Still OK with the testable, tested version?
[10:40] <anpok_> yes
[17:11] <alan_g> kdub: it's an old bug, but I think you landed some stuff that may be relevant. Could you take a look when you have an idle moment? https://bugs.launchpad.net/mir/+bug/1383745
[17:13] <kdub> alan_g, alright