[00:15] <RAOF> Wait. Is EGL actually choosing a 565 config somewhere? What could possibly possess it to do that?!
[03:42] <RAOF> And, uh, now I've somehow managed to get my laptop, that I'm not even running Mir on, in a state where GNOME won't turn the internal panel on?
[04:24] <RAOF> Aaargh. It's not stride, it's choosing the a tiling format it can't scanout of properly!
[04:25] <RAOF> s/tiling/modifier/, s/format//
[04:25] <RAOF> But I'm deliberately using the no modifiers codepath
[04:32] <RAOF> And, of course, deliberately asking for GBM_BO_USE_LINEAR makes it work on amdgpu and breaks nouveau. Superb.
[07:36] <RAOF> Huh. Is that why bypass was failing on the RPi 4?
[07:36] <RAOF> We were never releasing the GEM handles?
[07:45] <RAOF> Serendipitous discoveries while trying to work out why, why, amdgpu insists on allocating a buffer with layout incompatible with scanout while being asked for GBM_BO_USE_SCANOUT.
[07:50] <Saviq> Yay, CVE day \o/
[08:06] <RAOF> Oh, 6pm.
[08:17] <Saviq> RAOF (he/they) make sure to register for the Summit, if you haven't yet: https://events.canonical.com/event/31/
[08:18] <RAOF> Yup, I have done
 "Yay, CVE day \o/" <- Are you handling that? Or shouldI do some (or all)?
[08:47] <Saviq> I'm handling it. Muscle memory by now. Though I don't remember a stack like this one (it was libx11, so everything went)
[08:47] <alan_g[m]> Think about how much more we could automate.
[11:23] <Saviq> I should've said.... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/9847cbc1d68d3e30b99d3085c3f311fadac8554f>)
 "I should've said...." <- > <@saviq:matrix.org> I should've said.... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/9de1add62a0cf87a5f46dd0fb69a42652840e097>)
[16:27] <Saviq> Today:
[16:27] <Saviq> - got a slew of CVEs through (just mir-kiosk awaiting QA)
[16:27] <Saviq> - fixed up mir-test-tools jobs, split out performance tests into separate units
[16:27] <Saviq> alan_g W00t on Drag'n'drop!
[16:28] <alan_g[m]> Saviq: Let's talk about testing for it on Monday
[16:31] <alan_g[m]> Today:
[16:31] <alan_g[m]> * Realised that testing DnD needs test infrastructure we don't yet have
[16:31] <alan_g[m]> * Cleaned up the code some more and PR's the result
[16:51] <Saviq> Good weekend all 🌊