[07:04] <Saviq> G'day \o
[07:52] <RAOF> Ooooh! *That's* why the smoke tests fail here! But why is it trying to use the gbm renderer on X11? :blobfox_thinking: 
[07:59] <RAOF> Ah, but that should be for tomorrow.
[08:06] <RAOF> Today:
[08:06] <RAOF> * Reviewed the last WLCS PR lying around - [Sophie's popup dismissal ordering PR](https://github.com/MirServer/wlcs/pull/223). There are now no more outstanding WLCS PRs in an almost-mergable state 🎉
[08:06] <RAOF> * Hunted down why the smoke-tests are failing (on X11) in platform-API branch: it's incorrectly using the gbm rendering platform, rather than the X11 rendering platform
[08:06] <RAOF>   -
[08:07] <RAOF>  * Today:... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/09b091e1f1be5974e1dbb02dcfdb968ae7828e38>)
[08:07]  * RAOF needs to come up with a better name than DumbDisplayProvider, too.
[08:08] <RAOF> Now soliciting suggestions :)
[08:11] <Saviq> RAOF: > * <@chris:cooperteam.net> needs to come up with a better name than `DumbDisplayProvider`, too.
[08:11] <Saviq> Stub?
[08:16] <alan_g[m]> What does DumbDisplayProvider do? How does it differ from {,Foo}DisplayProvider?
[08:40] <RAOF> It provides CPU accessible framebuffer.
[08:40] <RAOF> s/framebuffer/framebuffers/
[08:41] <RAOF> Which, I guess, already provides a better name!
[08:41] <alan_g[m]> You're welcome!
[09:27] <alan_g[m]> FWIW we've done well landing stuff, this was a couple of pages when I opened it on Monday, now down to a more managable four items:
[09:27] <alan_g[m]> https://github.com/pulls?q=is%3Aopen+is%3Apr+org%3AMirServer+archived%3Afalse+draft%3Afalse
[11:12] <Saviq> Hey sophie, you said in Prague we don't support fractional scaling, but my experience with Miriway is that… we do, at least to some extent (I mean, if I use `scale: 1.5`, it's different to `scale: 1` and `scale: 2`). Or did you mean we're missing protocol support, for example?
[11:21] <alan_g[m]> <Saviq> "Hey sophie, you said in Prague..." <- It's (some of) the Wayland protocols that decided scale is integer: we need to work stuff through as clients should be able to pre-scale buffers for better rendering
[11:55] <alan_g[m]> <Eighth_Doctor> "alan_g, RAOF (he/they): do..." <- As Saviq said, Sophie did the work to prove a migration path with a few desktop components working on Wayland, but nothing meaning has happened since.
[11:56] <Eighth_Doctor> alan_g: how far along is Mir's Wayland protocol support?
[11:56] <Eighth_Doctor> vs wlroots or kwin, for example?
[11:57] <alan_g[m]> It depends on the protocol. For example xdg-shell:
[11:57] <alan_g[m]> https://wayland.app/protocols/xdg-shell#compositor-support
[11:59] <alan_g[m]> https://wayland.app/protocols/wayland#compositor-support
[12:21] <Saviq> alan_g does Alt not get stuck for you after Alt+Tab from time to time in Miriway?
[12:42] <Saviq> alan_g also, can you interact with X decorations? I can't drag or X.
[13:00] <alan_g[m]> I noticed "X" not working yesterday (not filed a bug yet). Can now confirm other controls and drag/resize don't work either.
[13:01] <Saviq> Filed 
[13:01] <Saviq> https://github.com/MirServer/mir/issues/2943
[13:02] <Saviq> https://github.com/MirServer/mir/issues/2583 also back as far I can tell
[13:06] <Saviq> Ah, but that last one is b/c we've not enabled primary selection by default
[13:07] <alan_g[m]> Saviq: It is a reasonable question whether we should (by default) for Miriway
[13:10] <Saviq> Looks like there's movement on the upstream bug at least:
[13:10] <Saviq> https://bugzilla.mozilla.org/show_bug.cgi?id=1791417
[13:19] <alan_g[m]> Saviq I'm going to bisect #2943 - is likely something stupidly simple
[14:58] <alan_g[m]> <Saviq> "https://github.com/MirServer/mir..." <- BFI fix proposed: https://github.com/MirServer/mir/pull/2945
[15:00] <Saviq> We should add Miriway to Mir CI
[15:02] <alan_g[m]> Feel free
[15:02] <Saviq> I know I know!
[15:11] <alan_g[m]> AN
[15:12] <alan_g[m]> Any thoughts on how to clean up $PATH without shopt -s extglob?
[15:17] <Saviq> sed?
[15:18] <alan_g[m]> <alan_g[m]> "Any thoughts on how to clean up..." <- Actually, I don't think it is needed anymore...
[16:01] <Saviq> A'it. Today:... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/951a2573baa8a02b15b34c28cd98d4242e79b72d>)
[16:22] <sophie-w> Conan Kudo: None of us have worked on MATE Wayland support for some time, and as far as I know no one from MATE has done much either. Some MATE things work in a Wayland session, many do not.
[16:24] <sophie-w> <Saviq> "Hey sophie, you said in Prague..." <- My memory is that someone else said we don't support fractional scaling and I said we do support it. in any case Alan is right that it kinda works currently and with support for more protocols we could probably make it look better.
[16:38] <alan_g[m]> Today:... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/d86fa913590af6a628782dfa17d1de4d1324ad30>)
[23:35] <RAOF> <sophie-w> "My memory is that someone else..." <- Specifically: there's an explicit [fractional scale](https://wayland.app/protocols/fractional-scale-v1) protocol, but also I understand some clients use [`wp_viewporter`](https://wayland.app/protocols/viewporter) to do a similar trick.
[23:36] <RAOF> So, we support fractional scaling, but we don't allow clients to submit fractionally-scaled buffers.