[07:48] <alan_g[m]> We forgot to update arch specific symbols! https://github.com/MirServer/mir/issues/2528
[08:06] <Saviq> That's a big diff… :confused:
[08:10] <alan_g[m]> Most of the diff is nonsense
[08:35] <alan_g[m]> Oh! When I eliminate the "spurious" diffs for mircore, nothing is left . What am I missing?!
[08:36] <Saviq> Yeah I didn't see any difference other than the Mir version…
[09:52] <alan_g[m]> Usually there's a or two line with another difference, and lots of noise.
[09:52] <Saviq> I compared all the +/- lines, haven't found a single one that would differ in any way other than the version string
[09:52] <alan_g[m]> Yeah, me too
[10:09] <alan_g[m]> I think it has to be the version. (We changed 2.8.0 to 2.9.0 without a changelog entry.) But don't see why only arm archs would be affected
[10:10] <Saviq> Let me try and reproduce
[10:11] <alan_g[m]> Coward! 😀 I was going to suggest pushing s/2.9.0/2.8.0/g to main and see
[10:58] <alan_g[m]> Saviq any joy reproducing?
[10:58] <Saviq> Joy, plenty. Reproducing, not so much.
[10:59] <Saviq> Just ran out of storage trying to build Mir on the ARM VM I had access to ;)
[11:00] <alan_g[m]> In that case I offer to push s/2.9.0/2.8.0/g to main. (What can go wrong)
[11:04] <Saviq> Fine with me
[11:05] -GitHub[m]:#mir-server- **[MirServer/mir]** AlanGriffiths closed [issue #2528](https://github.com/MirServer/mir/issues/2528): Broken .symbols on armhf
[11:05] <alan_g[m]> done
[11:33] <alan_g[m]> Well, that's a shorter diff...
[11:33] <alan_g[m]> https://launchpadlibrarian.net/616289811/buildlog_ubuntu-kinetic-amd64.mir_2.8.0+dev477-ge57d063ca1-0ubuntu22.10_BUILDING.txt.gz
[11:37] <alan_g[m]> ...and it makes sense. Just not sure why we've not seen it
[12:55] <Saviq> alan_g that's fallout from the above, is it?
[12:55] <Saviq> https://github.com/MirServer/mir/runs/7651316853?check_suite_focus=true#step:5:13529
[12:59] <Saviq> (this is a bors run for the two recently approved PRs)
[13:00] <alan_g[m]> Yes, I'm still using main to experiment
[13:05] <alan_g[m]> +dev478 is my second iteration at fixing. But results still pending
[13:13] <alan_g[m]> Saviq I guess we can delete impish packages from our PPAs now?
[13:13] <Saviq> Sure
[13:54] <alan_g[m]> First result in: armhf on 22.04 fixed
[14:30] <Saviq> Good. I have an armhf and arm64 environment for any future tests
[14:34] <alan_g[m]> Can you see where this failed: https://launchpad.net/~mir-team/+archive/ubuntu/dev/+build/24241856
[14:39] <Saviq> No, I think this is LP messing things up. I've seen it a handful of times yesterday
[14:39] <Saviq> Status: successful
[14:40] <alan_g[m]> Thanks for confirming
[14:42] <Saviq> I've asked and will ping LP team https://answers.launchpad.net/launchpad/+question/702651
[14:48] <alan_g[m]> OK, apart from LP weirdness I think normal service can be resumed. I'll retry the bors jobs...
[14:50] <Saviq> ACK
[15:07] <Saviq> In the ARM env I have I have XdgPopupTest failing - which I believe we've seen from time to time in CI and LP
[15:07] <Saviq> ./tests/xdg_popup.cpp:784: Failure
[15:07] <Saviq> Value of: manager->client.keyboard_focused_window()
[15:07] <Saviq> Expected: is equal to 0xe8b710
[15:07] <Saviq>   Actual: NULL (of type wl_surface*)
[15:19] <alan_g[m]> I've not investigated. But I suspect the test is racy
[15:30] <Saviq> Yeah, it's just that may mean we have an environment to reproduce at least.
[15:31] <alan_g[m]> Always a good first step
[15:31] <Saviq> How could adding coverage builds caused the wayland generator test to fail o.O
[15:31] <Saviq> https://github.com/MirServer/mir/runs/7654629196?check_suite_focus=true#step:9:153
[15:31] <Saviq> Is a question for tomorrow. I have to clock out a bit early today o/
[15:32] <alan_g[m]> Well, easy: the instrumentation does change something
[15:45] <alan_g[m]> Now for the fun! Spread wants ctor/dtor/vtab symbols for `mir::ShmFile` while the PPA doesn't. Something different in the build chain?!
[15:50] -GitHub[m]:#mir-server- **[MirServer/mir]** AlanGriffiths opened [pull request #2529](https://github.com/MirServer/mir/pull/2529): Testing this spread builds 
[15:50] -GitHub[m]:#mir-server-  
[15:50] -GitHub[m]:#mir-server- > Re-add symbol expectations that the PPA was complaining were missing
[16:14] <sophie-w> alan_g: did you intend to push symbol stuff directly to main?
[16:15] <alan_g[m]> Yes, we we trying to fix PPA builds
[16:17] <alan_g[m]> But now it seems that broke the spread builds. Am investigating
[16:31] -GitHub[m]:#mir-server- **[MirServer/mir]** AlanGriffiths merged [pull request #2529](https://github.com/MirServer/mir/pull/2529): Testing this spread builds
[16:32] <alan_g[m]> Spread fixed. Now to see what the PPA makes of it...
[16:49] <alan_g[m]> As feared: the PPA complains
[16:53] <alan_g[m]> OK, I'm getting too close to EOD to sort it out. Things should be working in CI, but with the PPA complaining about ctor/dtor/vtab symbols for mir::ShmFile.
[16:54] <alan_g[m]> No sure why the difference, but I'll be "on it" tomorrow
[16:58] <alan_g[m]> And that's it from me...
[17:16] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2371](https://github.com/MirServer/mir/pull/2371): Modernize surface observer
[17:46] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2504](https://github.com/MirServer/mir/pull/2504): Support zwlr_virtual_pointer_v1