[09:02] <Saviq> OK so that's a bit dum-dum, Snapcraft won't let you built with a `build-provider: ` that doesn't exist. But it will, with a channel that does. Except SnapD will still fail to install:... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/c7174987fe0112339e9a30980f2951bf6bee24de>)
[09:07] <alan_g[m]> Yeah, I've complained about it in the past.
[09:08] <alan_g[m]> You have to install manually
[09:34] <Saviq> But I think you at least could build with a snap not-in-stable, now you can't
[09:35] <Saviq> And then if you do use a channel, you can't install the built snap…
[09:35] <Saviq> So the only real option is to just omit the default provider bit
[09:36] <Saviq> (or well, release to stable)
[09:36] <alan_g[m]> Well, I've installed mesa-core22/edge and then I can install the consumer snaps built with .../edge
[09:37] <Saviq> OK yeah, you can install first, but that doesn't help that much
[09:37] <Saviq> Snapcraft should probably just let things through despite default-provider not being there at built time
[09:37] <alan_g[m]> +1
[09:44] <Saviq> Filed https://bugs.launchpad.net/snapcraft/+bug/2012057
[09:44] <Saviq> And will modify everything to default-provider: mesa-core22
[09:48] <Saviq> Ugh. But that will make them not build 🤦‍♂️
[09:49] <Saviq> Will leave it at /edge for now, then
[12:26] <alan_g[m]> Would it break anything to promote mesa-core22 to stable? (It doesn't need to be public)
[12:27] <Saviq> No, but it also wouldn't help much until we publish them under a single publisher (no auto-connection)
[12:27] <Saviq> But yeah at least it would install
[12:36] <Saviq> OK done, it's on stable now
[13:58] <alan_g[m]> Testing OSK I discovered that the keyboard appears at the top of the display. Doesn't seem to be the core22 changes though. Can you reproduce?
[13:58] <Saviq> Worked fine when I tried in Miriway, will try a real session
[14:01] <alan_g[m]> Seems to be a difference between Frame/edge and Frame/stable
[14:03] <alan_g[m]> Frame/beta also shows it wrong
[14:04] <alan_g[m]> And candidate!
[14:05] <alan_g[m]> I'm using it with frame-it (with osk=enable)
[14:11] <Saviq> Yeah, that I can reproduce
[14:27] <Saviq> Will leave it with you, digging into our tests crashing on Pi4 with core22
[15:14] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** AlanGriffiths opened [pull request #132](https://github.com/MirServer/ubuntu-frame/pull/132): Fix OSK... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/5ea577b94cc2dd68afcedd04a16ac3727e31be6d>)
[16:07] -GitHub[m]:#mir-server- **[MirServer/mir]** Saviq opened [issue #2870](https://github.com/MirServer/mir/issues/2870): `mir_performance_tests` crashing on armhf on 22.04... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/283a9e076ebb3d35bd9716ba5ce8eae53e1b27f9>)
[16:17] <alan_g[m]> So the question is, do we fix Frame now, or just roll it into the Mir 2.13 update
[16:26] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** bors[bot] edited [pull request #132](https://github.com/MirServer/ubuntu-frame/pull/132): [Merged by Bors] - Fix OSK
[16:26] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** bors[bot] closed [pull request #132](https://github.com/MirServer/ubuntu-frame/pull/132): [Merged by Bors] - Fix OSK
[16:30] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** Saviq assigned  to [pull request #132](https://github.com/MirServer/ubuntu-frame/pull/132): [Merged by Bors] - Fix OSK
[16:37] -GitHub[m]:#mir-server- **[MirServer/wlcs]** wmww requested a review from RAOF for [pull request #277](https://github.com/MirServer/wlcs/pull/277): Test XDG shell (stable) version 5
[17:43] <Saviq> sophie https://github.com/MirServer/mir-ci/actions/runs/4264568386/jobs/7814456136?pr=5#step:8:2342 is now running a locally-built egmde in a loop, you should be able to SSH with:
[17:43] <Saviq> ```plain
[17:43] <Saviq> ssh QWqADaOdUoyPynH99AXU:MWNlNmQ0OGEudm0udXB0ZXJtLmludGVybmFsOjIyMjI=@uptermd.upterm.dev
[17:43] <Saviq> ```
[17:44] <Saviq> But of course it's now been running for a few minutes and no crash…
[17:48] <Saviq> We may just put egmde to rest over this.
[17:53] <sophie-w> Saviq: uhh, do I need keys for that?
[17:53] <Saviq> I imported your LP ones?
[17:53] <sophie-w> hmm, didn't work
[17:54] <Saviq> Ah, meh, they'd need to be imported on the server side of this… too late
[17:55] <Saviq> Anyway, nothing to see there as it didn't segfault…
[17:57] <sophie-w> ok makes sense, confirmed that my launchpad pubkey is up to date and correct
[18:01] <Saviq> There's one important difference here - egmde is classic, so it's running with the host's (22.04) bits around.
[18:08] <Saviq> But pulling the host back to 20.04 doesn't help
[18:15] <Saviq> Trying one last thing - building on 20.04 rather than 22.04
[18:51] <Saviq> OK that worked
[18:51] <Saviq> Will get an env you can SSH into
[19:39] <Saviq> sophie can you `ssh KcZqYeJT6gAFKoZPWlYc:MWNlNmQ0OGEudm0udXB0ZXJtLmludGVybmFsOjIyMjI=@uptermd.upterm.dev`?
[19:40] <Saviq> You can ;)
[19:41] <sophie-w> I can
[19:43] <Saviq> OK I'll get it to the point where the issue can be reproduced and let you know
[20:15] <Saviq> sophie OK it's all yours, looping over tests now - of course it's now happily chugging along.
[20:19] <sophie-w> Thanks! have a good evening
[20:19] <sophie-w> and weekend