[06:38] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** Saviq marked [pull request #117](https://github.com/MirServer/ubuntu-frame/pull/117): snap: migrate to `base: core22` as ready for review
[07:54] <Saviq> Mornang o/
[08:42]  * Saviq uploaded an image: (735KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/ygVuppZBaIBZiLoTSbxwWcXf/image.png >
[08:42] <Saviq> alan_g hey, scummvm in confined-shell launching like this is the SDL fullscreen problem?
[08:42] <Saviq> It fixes itself on hover
[08:46] <alan_g[m]> I would guess so. Should I dig into it?
[08:47] <Saviq> No, just confirming we've seen this. Can't find an issue logged, though, can you?
[09:02] <alan_g[m]> Try another SDL2 app? mir-kiosk-neverputt?
[09:03] <Saviq> Yeah that one's fine
[09:05] <alan_g[m]> Now I'm getting curious (but it does suggest the problem is on the client side)
[09:09] <alan_g[m]> I see the same problem (unsurprisingly) on normal Miriway
[09:11] <alan_g[m]> > WARNING: SDL mixer output buffer size 512 differs from desired: 1024!
[09:14] <alan_g[m]> And no, I don't see an issue
[09:37] <alan_g[m]> I guess "SDL mixer output" is sound though
[09:46] <alan_g[m]> In response to Mir configuring the window to 2560x1440 scummvm is sending a 640x400 buffer and then "ack"ing the configure. Only after later events does it send a 2560x1440 buffer
[09:48] <Saviq> Wonder what GNOME does differently?
[09:49] <alan_g[m]> It doesn't. Same happens on GNOME
[09:59] <Saviq> Hmm not here, maybe b/c two displays?
[10:00] <alan_g[m]> I've two displays on 22.04 and one on 23.04 in both cases GNOME shows the partial window
[10:01] <Saviq> Fullscreen all around for me, one or two displays on 22.10 🤷
[10:01] <Saviq> But if you can reproduce on GNOME I say: not our problem
[10:01] <Saviq> IOW: bug with ScummVM
[12:32] <Saviq> Wop. Got mir-kiosk-kodi on Nvidia 👯
[12:32] <Saviq> Just need to get VDPAU (¿or VAAPI through VDPAU?) working now…
[12:32] <alan_g[m]> Nice!
[12:33] <Saviq> Could do, yes
[12:33] <alan_g[m]> Maybe it needs to become frame-kodi for core22?
[12:48] <Saviq> Hmm maybe VDPAU isn't that interesting after all… https://kodi.wiki/view/Supported_hardware#Desktop
[12:48] <Saviq> Would have to check if it works on Classic in the first place
[12:56] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame-vnc]** Saviq opened [pull request #7](https://github.com/MirServer/ubuntu-frame-vnc/pull/7): snap: move to graphics-core22
[12:56] -GitHub[m]:#mir-server-  
[12:56] -GitHub[m]:#mir-server- > Also bring `scripts/` in, rather than staging from the deprecated snap-launch.
[14:57] <Saviq> sophie I can see you tried to [restart the egmde runs](https://github.com/MirServer/mir-ci/actions/runs/4264568386) in mir-ci - were you looking for a way to SSH in there?
[14:57] <Saviq> If so, you need to check the "Enable debug logging" bit (I know, it's not intuitive, but it's the only UI piece you can change on restart)
[15:10] <Saviq> I did get them to fail locally with timeout 3600 sh -c 'while env MIR_TEST_SERVER=egmde xvfb-run pytest -n8; do true; done' after a bunch of runs
[15:10] <Saviq> (the -n8 is from pytest-xdist)
[15:10] <Saviq> I'll be back later o/