[04:41] <RAOF> Oh, no. I know what the `mmap` armhf bug is 🤦‍♀️
[05:51] -GitHub[m]:#mir-server- **[MirServer/mir]** RAOF opened [pull request #2549](https://github.com/MirServer/mir/pull/2549): mtf/mmap_wrapper: Pick correct mmap symbol on 32bit platforms... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/46af90f208ef1f1cd3cd54aab66f871e49351e34)
[05:59] -GitHub[m]:#mir-server- **[MirServer/mir]** Saviq edited [pull request #2549](https://github.com/MirServer/mir/pull/2549): mtf/mmap_wrapper: Pick correct mmap symbol on 32bit platforms
[06:01] <Saviq> RAOF (he/they) I retargeted ^ on @sophie's `mmap-virtual-keymap` to help confirm your fix, can you rebase so that the changes look saner?
[06:01] <RAOF> Saviq: Sure.
[06:53] <RAOF> TFW you go to check on CI just to find that you forgot to press the hardware-SSH-key button and so the branch didn't actually push yet 🤦‍♀️
[08:37] -GitHub[m]:#mir-server- **[MirServer/mir]** AlanGriffiths marked [pull request #2548](https://github.com/MirServer/mir/pull/2548): Don't destroy CompositingFunctors that are not ready as ready for review
[08:41] <alan_g[m]> Saviq headspace to review ^?
[08:42] <Saviq> ACK
[08:46] -GitHub[m]:#mir-server- **[MirServer/mir]** AlanGriffiths merged [pull request #2549](https://github.com/MirServer/mir/pull/2549): mtf/mmap_wrapper: Pick correct mmap symbol on 32bit platforms
[09:03] <Saviq> alan_g so IIUC by doing the current_exception / rethrow dance, you ensure the lifetime of the exception thrown?
[09:04] <alan_g[m]> Saviq: No, I ensure there's no thread using "other" objects when they are deleted
[09:04] <alan_g[m]> The logic around there could do with a proper cleanup. But that's a minimal fix
[09:06] <Saviq> ACK, tiny nitpick there
[09:07] <alan_g[m]> Why?
[09:08] -GitHub[m]:#mir-server- **[MirServer/mir]** Saviq marked [pull request #2540](https://github.com/MirServer/mir/pull/2540): mmap keymap fd instead of reading from it as ready for review
[10:45] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2548](https://github.com/MirServer/mir/pull/2548): Don't destroy CompositingFunctors that are not ready
[10:45] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] closed [issue #2547](https://github.com/MirServer/mir/issues/2547): MultiThreadedCompositor.when_compositor_thread_fails_start_reports_error is flaky
[13:59] <alan_g[m]> Saviq does `display-quirks=trykms:device:virtio_gpu` work as a new quiirk option?
[14:02] <Saviq> `trykms` does not ring the same as "disable modeset probe", not too broad?
[14:03] <alan_g[m]> You'd prefer `display-quirks=disable-modeset-proble:virtio_gpu`?
[14:04] <Saviq> Not really, but it is more specific…
[14:04] <alan_g[m]> I'm struggling to find a good incantation
[14:05] <alan_g[m]> `display-quirks=override-modeset-probe:virtio_gpu`
[14:06] <Saviq> kms-probe?
[14:06] <Saviq> disable-kms-probe I mean
[14:07] <alan_g[m]> s/override/disable/, s/modeset/kms/
[14:08] <alan_g[m]> Huh, that shouldn't have been an edit. Lousy UIU
[14:09] <alan_g[m]> So we have: `display-quirks=disable-kms-probe:virtio_gpu`
[14:13] <Saviq> You had `:device` before, is that not needed after all?
[14:22] <alan_g[m]> It should have been "driver" anyway, but I don't think "devnode" makes sense here
[15:47] -GitHub[m]:#mir-server- **[MirServer/mir]** AlanGriffiths opened [pull request #2550](https://github.com/MirServer/mir/pull/2550): Add a new driver quirk to disable KMS modeset probe
[15:47] -GitHub[m]:#mir-server-  
[15:47] -GitHub[m]:#mir-server- > With an implicit `display-quirks=disable-kms-probe:virtio_gpu` as we need it for virtio
[15:52] <alan_g[m]> Saviq here's the quirk you were after. 😁
[15:53] <Saviq> 👍
[16:31] -GitHub[m]:#mir-server- **[MirServer/mir]** Saviq self-requested a review for [pull request #2550](https://github.com/MirServer/mir/pull/2550): Add a new driver quirk to disable KMS modeset probe
[16:33] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2540](https://github.com/MirServer/mir/pull/2540): mmap keymap fd instead of reading from it
[16:42] <Saviq> Wop. Will test VNC input first thing tomorrow (snap should build soon to `edge`)
[16:42] <Saviq> And EOW o/
[16:43] <alan_g[m]> Have a good weekend (and don't work too hard!!)
[17:01] <alan_g[m]> ...and EOW for me too o/
[17:31] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2544](https://github.com/MirServer/mir/pull/2544): [miral] Improvement to ExternalClientLauncher
[18:50] <Saviq> D'uh. Tomorrow == Tuesday, since we have a day off here on Monday ;P
[19:43] -GitHub[m]:#mir-server- **[MirServer/mir]** wmww opened [pull request #2551](https://github.com/MirServer/mir/pull/2551): Drop (accidental?) extern "C++" in miaral symbols
[19:43] -GitHub[m]:#mir-server-  
[19:43] -GitHub[m]:#mir-server- > Not sure how that slipped through, at least on LLD it prevents Mir from building.
[20:12] -GitHub[m]:#mir-server- **[MirServer/mir]** wmww opened [pull request #2552](https://github.com/MirServer/mir/pull/2552): Expose miral::Zone::id()
[20:12] -GitHub[m]:#mir-server-  
[20:12] -GitHub[m]:#mir-server- > Not having this has been a recurring annoyance in writing shells. With the ID a shell can have a hashmap of zones, store an associated zone with a window, etc. Without it identifying zones is much less convenient and efficient.
[20:32] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** wmww opened [pull request #83](https://github.com/MirServer/ubuntu-frame/pull/83): Enable input protocols for VNC snap
[20:32] -GitHub[m]:#mir-server-  
[20:32] -GitHub[m]:#mir-server- > Possibly now that virtual pointer and seat v8 support have landed in Mir. The `ubuntu-frame-vnc` snap still doesn't seem to be trying input, but it's not obvious where that's disabled.
[20:55] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2546](https://github.com/MirServer/mir/pull/2546): [miral] Tidy up event filtering API
[22:21] <sophie-w> grayson-g: if you're still around, quick review on https://github.com/MirServer/mir/pull/2551?
 "grayson-g: if you're still..." <- I wonder why the tests are passing because `miral/symbols.map` was updated in the last PR that got merged. There were more things added to that `extern C++` block.
[22:31] <sophie-w> Maybe the linker CI uses ignores the cruft, but lld errors?
[22:32] -GitHub[m]:#mir-server- **[MirServer/mir]** wmww opened [pull request #2553](https://github.com/MirServer/mir/pull/2553): Add virtual pointer protocol name to miral
[22:32] -GitHub[m]:#mir-server-  
[22:32] -GitHub[m]:#mir-server- > Forgot this in original PR
[22:32] <grayson-g[m]> It should be out of date if you pull the newest `main`. Is `bors`ing this going to remove the new symbols?
[22:33] <grayson-g[m]> https://github.com/MirServer/mir/pull/2546/files#diff-b513d8e03756a3b8ea2c62ba4631ab7a500913d9b5b4253a89848811e63a99a5 referring to this
[22:44] <sophie-w> I think you can just merge it, it seems to cleanly merge
[22:45] <sophie-w> I locally rebased without conflicts and pushed, I guess git is smart enough to figure it out
[22:45] <sophie-w> Oop, my push cancelled your merge
[22:45] <grayson-g[m]> All good!
[22:46] <sophie-w> Try again?
[23:36] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2551](https://github.com/MirServer/mir/pull/2551): Drop (accidental?) extern "C++" in miaral symbols