[00:10] -GitHub[m]:#mir-server- **[MirServer/mir]** wmww edited [pull request #2495](https://github.com/MirServer/mir/pull/2495): Attempt to refactor geometry types without breaking MirAL ABI [01:04] "And https://github.com/MirServer..." <- Can be dropped [07:50] Ah. I see CI is on fire with the snap store being unavailable. [08:05] Yeah it will be quite unhappy… [11:52] sophie hey, not sure what's expected after ↑, but ubuntu-frame-vnc still reports 100% damage. That said, it does throttle frames down to 0 when nothing changes. I suppose that's expected when there's only a single full-screen app? [14:31] @Saviq yeah, we don't yet track damage areas of a surface, we treat a whole surface as damaged every time it gets a new frame [14:33] But if you had multiple displays with different apps, or split screen, damage would be useful (which I think is relevant, because you really don't want to capture two full displays worth of content twice as much as needed) [14:35] I suppose actually sending the damage to the client instead of just using it to know when to capture isn't getting us that much, but by the time everything else was wired up it was easy to implement [14:36] Tracking surface-level damage would be a useful thing not only for screencopy, but also for some platforms (according to RAOF), but I don't see that as a priority [14:42] Tracking damage is an optimisation. We will need it for various things, but those things need to first work. [16:11] RE https://github.com/MirServer/mir/pull/2494#discussion_r908256638: I fail to see how a visible ImmediateExecutor type would be useful (aside from maybe using dynamic_cast<>ing to check if things are it, which we probably want to avoid). Since there's only one possible implementation of ImmediateExecutor, anything that would take an ImmediateExecutor could simply take nothing and hard-code in the one to use. [16:27] Good night o/ [17:45] -GitHub[m]:#mir-server- **[MirServer/mir]** wmww opened [pull request #2496](https://github.com/MirServer/mir/pull/2496): Refactor SurfaceObserver types to be compatible with new observer system [17:45] -GitHub[m]:#mir-server- [17:45] -GitHub[m]:#mir-server- > Because the modern observer system does not always send notifications on the same thread, it needs to be given types it can hold ownership of. This PR changes types used by SurfaceObserver to work. All methods still take a raw surface pointer, which is generally not used and should be removed, but that is out of scope here. [17:50] -GitHub[m]:#mir-server- **[MirServer/mir]** wmww opened [pull request #2497](https://github.com/MirServer/mir/pull/2497): Use new MOCK_METHOD macro in MockSceneSession [17:50] -GitHub[m]:#mir-server- **[MirServer/mir]** wmww edited [pull request #2497](https://github.com/MirServer/mir/pull/2497): Use new MOCK_METHOD macro in MockSceneSession [17:51] -GitHub[m]:#mir-server- **[MirServer/mir]** wmww closed [pull request #2493](https://github.com/MirServer/mir/pull/2493): AbstractShell: refactor confinement [17:51] -GitHub[m]:#mir-server- **[MirServer/mir]** wmww reopened [pull request #2493](https://github.com/MirServer/mir/pull/2493): AbstractShell: refactor confinement [18:48] grayson-g: easy review: https://github.com/MirServer/mir/pull/2497 [18:50] gmock has added a new, improved syntax for mock functions since most of our mocks were written. Not only does it prevent you from having to specify the number of arguments in the macro name, it allows marking methods `override` (which prevents us accidentally failing to update/remove them when the base class changes) [18:50] on it [19:25] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2497](https://github.com/MirServer/mir/pull/2497): Use new MOCK_METHOD macro in MockSceneSession [19:42] -GitHub[m]:#mir-server- **[MirServer/mir]** wmww opened [pull request #2498](https://github.com/MirServer/mir/pull/2498): Allow --add-wayland-extenions all and clean up help message... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/5332a8fadc54ce441f20b7339aaa8888288b492d) [19:45] GitHub -> Matrix bot seems needlessly verbose to me. Would it be possible to: [19:45] - Not include full issue/PR text [19:45] - Not include issue/PR edited/closed (only opened)? [19:45] I would look into that myself if I had any idea where it is hosted/configured [19:45] * GitHub -> Matrix bot seems needlessly verbose to me. Would it be possible to: [19:45] - Not include full issue/PR text [19:45] - Not include issue/PR edited/closed (only opened)? [19:45] I would look into that myself if I had any idea where it is hosted/configured [19:47] -GitHub[m]:#mir-server- **[MirServer/mir]** wmww edited [pull request #2498](https://github.com/MirServer/mir/pull/2498): Allow --add-wayland-extenions all and clean up help message [19:55] > <@wmww:matrix.org> GitHub -> Matrix bot seems needlessly verbose to me. Would it be possible to:... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/dba25b76e8ab92823b1755622801e79e29e13e36) [19:57] sophie can you see https://github.com/MirServer/mir/settings/hooks/360169876 ? [19:59] If you can, you'll see that there's very little that's enabled, but I agree it's too verbose still [20:21] 404 :( [20:30] "404 :(" <- How about now? [20:32] https://github.com/MirServer/mir/settings/hooks/360169876 still 404s for me [20:32] sophie-w: Should be there now. Apparently that's "sensitive or destructive" [20:32] (it's fine, I believe you've turned off what you can) [20:32] Thanks, now I see it [21:13] -GitHub[m]:#mir-server- **[MirServer/mir]** bors[bot] merged [pull request #2496](https://github.com/MirServer/mir/pull/2496): Refactor SurfaceObserver types to be compatible with new observer system [21:40] -GitHub[m]:#mir-server- **[MirServer/mir]** wmww opened [issue #2499](https://github.com/MirServer/mir/issues/2499): Upgrade to wl_seat v7 [21:40] -GitHub[m]:#mir-server- [21:40] -GitHub[m]:#mir-server- > We currently expose v6, wayvnc wants v7. [23:19] "Tracking surface-level damage..." <- Specifically: you can pass damaged areas through to KMS, and some hardware (particularly: USB display adapters) can use that to optimise things.