03:26 alguien: Would anyone know why this is happening when watching vids? https://gemoo.com/tools/upload-video/share/613684546763874304?codeId=vzxa1on3N6an0&card=613684542909308928
03:40 wlb: wayland Issue #438 opened by Kirill Chibisov (kchibisov) Should `wl_display_connect` try `WAYLAND_DISPLAY` when `WAYLAND_SOCKET` is set but empty https://gitlab.freedesktop.org/wayland/wayland/-/issues/438 [IPC library]
08:04 wlb: weston Merge request !1452 opened by diegonieto (diegonieto) compositor/main: warn pipewire-output without mode https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1452
09:04 Ermine: Sorry for bringing this topic up again, but is there written-down reason for not supporting window placement in wayland?
09:05 Ermine: I guess I should read a discussion in the MR 247
09:16 pq: Ermine, Wayland in general targets things where there is no well-defined 2D coordinate system, and/or displays are not axis-aligned rectangles. Another reason is that a client cannot know of everything on the display, so it cannot make the best window positioning decisions.
09:16 pq: I'm not sure if such a write-up has been saved anywhere.
09:19 pq: Also the potential for window placement policy fighting between an app and a compositor.
09:21 pq: note, this is about global or per-output x,y coordinate based positioning. There are other ways to design positioning such that it doesn't clash with Wayland design principles that much.
09:27 Ermine: thank you!
09:51 wlb: wayland/main: Thomas Lukaszewicz * Mitigate UAF crashes due to iteration over freed wl_resources https://gitlab.freedesktop.org/wayland/wayland/commit/d275bc7f84f1 src/wayland-server.c tests/resources-test.c
09:51 wlb: wayland Merge request !360 merged \o/ (Mitigate UAF crashes due to iteration over freed wl_resources https://gitlab.freedesktop.org/wayland/wayland/-/merge_requests/360)
11:52 wlb: weston Issue #874 closed \o/ ([DRM] Alpha Background Support for the Kiosk Shell https://gitlab.freedesktop.org/wayland/weston/-/issues/874)
12:24 pq: What's the problem with eventfd? Can't one unregister it by simply removing it from epoll and close()'ing it?
12:24 pq: re: linux-drm-syncobj-v1
14:07 wlb: wayland Issue #438 closed \o/ (Should `wl_display_connect` try `WAYLAND_DISPLAY` when `WAYLAND_SOCKET` is set but empty https://gitlab.freedesktop.org/wayland/wayland/-/issues/438)
14:19 MrCooper: pq: it's about the DRM_IOCTL_SYNCOBJ_EVENTFD ioctl, it allocates a struct eventfd_ctx which keeps the eventfd alive until it's signaled (or the syncobj is destroyed)
14:20 pq: MrCooper, why is close(2) not hooked up to release it?
14:21 MrCooper: I don't know of any mechanism for that
14:23 pq: how can a file description remain alive after the process has no fds referring to it anymore? It that normal?
14:24 MrCooper: actually I'm not sure offhand the eventfd file description itself stays alive, the eventfd_ctx does though
14:25 pq: Seems very strange to me. The UAPI side is intuitively obvious how to release them.
14:25 emersion:doesn't really care about this stuff
14:26 pq: Are they accounted to the compositor process, or are they just generic kernel memory allocations not accounted to anyone like many things in DRM are AFAIU?
14:28 pq: I mean, all that compositors care about is to not exhaust their fd or memory quota. Anything else is the kernel's own problem.
14:32 pq: IOW, does it matter if the eventfd_ctx is not freed timely?
14:42 MrCooper: AFAICT it doesn't
14:46 MrCooper: just trying to get to the bottom of that alluded issue
16:08 dubiousness: pq: do we know how 3D displays handle positioning? Things like VR headsets?
16:09 dubiousness: I’d imagine it’s a relational X,Y,Z coordinate system, though I must admit I haven’t looked into it
16:30 bl4ckb0ne: dubiousness: for 2d clients in 3d space?
16:46 dubiousness: bl4ckb0ne: yes
16:47 bl4ckb0ne: in wxrc i just let the compositor handle the Z axis and treat x and y as a big screen
16:47 emersion: but, each toplevel tree has its own 2d space
16:47 bl4ckb0ne: handle as in "theres a keybinding to bring the toplevel closer or further away"
16:48 bl4ckb0ne: yes, its just a plane
19:34 wlb: wayland Issue #439 opened by Julian Orth (mahkoh) wl_surface::set_cursor with singleton cursor surface https://gitlab.freedesktop.org/wayland/wayland/-/issues/439
22:49 wlb: weston Merge request !1453 opened by Leandro Ribeiro (leandrohrb) Test suite changes related to ICC profiles https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1453 [Colour management], [Testing]