[06:33] <Saviq> Morn'
[06:35] <RAOF> 👋
[07:30] <Saviq> sophie on damage reporting… I wonder how important that is when hardware accelerated encoding is used… In the best case scenario, there should be zero copy involved, and I imagine the hardware encoder is quite good at "detecting damage"… like, that's most of what video compression is about?
[07:45] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2470](https://github.com/MirServer/mir/pull/2470): Send changed area to frame posted notification
[07:45] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] edited [pull request #2481](https://github.com/MirServer/mir/pull/2481): Refactor scene and surface change notification classes
[08:16] -GitHub[m]:#mir-server- **[MirServer/mir]** RAOF opened [pull request #2482](https://github.com/MirServer/mir/pull/2482): Surface: Drop some cursor_stream detritus left over from mirclient
[08:36] <alan_g[m]> Saviq shouldn't a successful "close snaps" close the snap branch? https://github.com/MirServer/mir-kiosk-kodi/pull/31#issuecomment-1160142878
[08:37] <Saviq> > Channel 'edge/pr31' for 'mir-kiosk-kodi' is now closed
[08:38] <Saviq> So. Hmm.
[08:39] <alan_g[m]> Oh, there's no version there
[08:39] <Saviq> Yeah it is closed
[08:39] <Saviq> The branch will be garbage collected after some time
[08:39] <Saviq> And if you refresh / install, you'll get `edge`
[08:39] <alan_g[m]> 🎆
[08:46] <alan_g[m]> Saviq is this intentionally empty? https://github.com/MirServer/mir-kiosk-x11-snap/pull/15
[08:47] <Saviq> No, fixing
[08:50] <Saviq> alan_g fixed, I confused GH before
[08:52] <alan_g[m]> There does seem to be a lot of repetition
[08:57] <Saviq> Less now, but a lot still, indeed.
[09:00] <Saviq> I don't think there's a way to reduce it much more. Each repo will need a workflow one way or another. Maybe I can put better defaults in the action so unless there's something special about a particular repo (like `snap: mir-kiosk-x11`), the defaults would do
[09:10] <alan_g[m]> Saviq this needs fixing https://github.com/MirServer/ubuntu-frame-vnc/pull/1#issuecomment-1160155235
 "sophie on damage reporting… I..." <- Does VNC have good compression? My impression was damage would greatly help it but I haven't measured that
[09:10] <Saviq> > Saviq this needs fixing https://github.com/MirServer/ubuntu-frame-vnc/pull/1#issuecomment-1160155235
[09:10] <Saviq> Yeah, I'll just drop the PR… I'm working on that snap right now
[09:10] <Saviq> > Does VNC have good compression? My impression was damage would greatly help it but I haven't measured that
[09:10] <Saviq> It can do h264, provided both the client and server can do it. So it still may be worhtwhile for jpeg (I _believe_ the default compression)
[09:10] <Saviq> And well, even with h264 it may be worthwhile if we can't achieve zero-copy
[09:12] <alan_g[m]> sophie-w: I'd guess that it depends on the actual path taken. It may not help the "best case" but it should help others.
[09:12] <Saviq> Indeed
[09:39] <alan_g[m]> Does this also need updating for the snapcraft changes? https://github.com/MirServer/ubuntu-frame/pull/44
 "Does this also need updating for..." <- Yeah will rebase them all, to but also probably make what's using it merge tips of the relevant projects
[09:46] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2473](https://github.com/MirServer/mir/pull/2473): ci: add armhf builds
[10:50] <alan_g[m]> Saviq easy review: https://github.com/MirServer/mircade/pull/10
[10:52] <RAOF> Best case, sending damage dramatically reduces the computational complexity of whatever encoding you're doing, and video encoders hate encoding the sort of super-high-frequency data that desktops consist of.
[10:53] <Saviq> Right, but that assumes the encoder can actually consume damage data, which I don't think will be the case for a hardware encoder?
[10:54] <Saviq> I may also be wrong ¯_(ツ)_/¯
[10:56] <Saviq> Like, I mean, for what the encoders get used to, damage data will rarely be available… one of the things that they need to find is the damage, so I would be surprised if there was API to supply damage from the outside.
[11:10] <Saviq> > So looking into abigail (for ABI stability reasons) there are two options:
[11:10] <Saviq> alan_g in case you missed this ^, what would be your preference?
[11:11] <Saviq> > <@saviq:matrix.org> So looking into abigail (for ABI stability reasons) there are two options:
[11:11] <Saviq> > 1. we compare deb packages (so probably grab the ones from Ubuntu devel) against what we build in CI
[11:11] <Saviq> > 2. we check in an XML representation and compare that against the .so we build
[11:11] <Saviq> Now with a better reference
[11:13] <alan_g[m]> I think 1 as stated has a problem as soon as we've released an ABI change to ppa:mir-team/release (should that be the reference?)
[11:14] <Saviq> It could be, yes. There should be very little delay between release and devel, but you're right, during freezes it could mess with us
[11:19] <alan_g[m]> I've a mild preference for not relying on pulling binaries off the internet though. Having a source of truth contemporaneous with the code could be useful working with historic releases
[11:35] <Saviq> Yeah the arguments for looking at the PPA are, AFAICT:
[11:35] <Saviq> - not having to maintain another ~symbols file
[11:35] <Saviq> - it's verifying the real thing
[11:35] <Saviq> Not that they're arguments enough
[11:39] <alan_g[m]> Hmm. More "random" wlcs test failures. Time to investigate...
[11:43] <Saviq> I'm inclined to blame qemu for this. They're all arm* failures, which run emulated
[11:44] <alan_g[m]> Maybe, but that could also simply be exposing an underlying problem
[11:45] <Saviq> I'll try and get them to fail on hardware
[11:47] <alan_g[m]> My usual tactic is to run the test in a loop until it fails. Even on different arch that often exposes a fail
[11:52] <Saviq> Yup, I sometimes add a `stress …` for good measure
[11:52] <alan_g[m]> Me too
[12:22]  * Saviq uploaded a video: (3724KiB) < https://libera.ems.host/_matrix/media/r0/download/matrix.org/pMvyGHDytCDONjUddtwwkvUU/Nagranie%20ekranu%20z%2020.06.2022%2014%3A20%3A24.webm >
[12:22] <Saviq> Hmm did we expect this to work yet? :D
[12:25] <Saviq> I thought there was something missing to get more than one frame from screencopy…
[12:39] -GitHub[m]:#mir-server- **[MirServer/wlcs]** vberger opened [issue #237](https://github.com/MirServer/wlcs/issues/237): xdg_shell_unstable_v6 is currently not optional in the test suite... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/e511bc8e18574f54667e3b7f4f551476956ed515)
[12:45] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2481](https://github.com/MirServer/mir/pull/2481): Refactor scene and surface change notification classes
[12:45] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2478](https://github.com/MirServer/mir/pull/2478): github: use canonical/actions/build-snap
[13:13] <Saviq> > Nagranie ekranu z 20.06.2022 14:20:24.webm
[13:13] <Saviq> This now available with ubuntu-frame/edge and ubuntu-frame-vnc/edge, the VNC server can be started with `ubuntu-frame-vnc --disable-input` (for now)
[13:16] <Saviq> alan_g not sure if it's clear what the Frame CfT is about, I read "Nvidia based desktop" as us supporting Nvidia in Frame.
[13:16] <Saviq> Maybe "X11 desktop"?
[13:17] <alan_g[m]> I'll update
[13:19] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2482](https://github.com/MirServer/mir/pull/2482): Surface: Drop some cursor_stream detritus left over from mirclient
[13:34] <Saviq> Today is bad infra day…
[13:38] <alan_g[m]> Hot weather in Europe?
[14:24] <Saviq> > I'll try and get them to fail on hardware
[14:24] <Saviq> No dice, been running well-stressed for a half hour now. Will try under qemu
[14:34] <alan_g[m]> Yeah, been warming up an old laptop since before lunch with no issues. (Always nicer to have a reproducer without too many hoops)
[14:47] <Saviq> OK good evening all o/
[14:49] <alan_g[m]> 👋
[14:49] -GitHub[m]:#mir-server- **[MirServer/wlcs]** vberger opened [pull request #238](https://github.com/MirServer/wlcs/pull/238): Make zxdg_shell_v6 ExpectedlyNotSupported if not in supported_extensions... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/304596e70b2f628df5cfff884d2d7bcf0430f507)
[15:04] -GitHub[m]:#mir-server- **[MirServer/wlcs]** bors[bot] merged [pull request #238](https://github.com/MirServer/wlcs/pull/238): Make zxdg_shell_v6 ExpectedlyNotSupported if not in supported_extensions
[15:04] -GitHub[m]:#mir-server- **[MirServer/wlcs]** bors[bot] closed [issue #237](https://github.com/MirServer/wlcs/issues/237): xdg_shell_unstable_v6 is currently not optional in the test suite
[17:21] -GitHub[m]:#mir-server- **[MirServer/wlcs]** vberger opened [pull request #239](https://github.com/MirServer/wlcs/pull/239): Destroy xdg_toplevel before xdg_surface on cleanup
[17:21] -GitHub[m]:#mir-server-  
[17:21] -GitHub[m]:#mir-server- > This is mandated by the protocol.
[17:36] -GitHub[m]:#mir-server- **[MirServer/wlcs]** vberger opened [issue #240](https://github.com/MirServer/wlcs/issues/240): WLCS relies on implicit focus on creation... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/c444f2f9dfe1e87e3362ba69f96744341b97592e)
[17:37] -GitHub[m]:#mir-server- **[MirServer/wlcs]** vberger edited [issue #240](https://github.com/MirServer/wlcs/issues/240): WLCS relies on implicit focus on creation
[17:52] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** Saviq opened [issue #75](https://github.com/MirServer/ubuntu-frame/issues/75): Is there any way to run chromium on Ubuntu Core? I've just installed chromium snap on a device running Ubuntu Core and I wasn't able to run it.... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/3cfe6e1123f586b72e3cab264a414c22a9edbcf2)
[17:54] -GitHub[m]:#mir-server- **[MirServer/ubuntu-frame]** Saviq edited [issue #75](https://github.com/MirServer/ubuntu-frame/issues/75): Is there any way to run chromium on Ubuntu Core?