08:29 daniels: riteo: definitely directly by clients as well, e.g. media players
08:53 pq: kennylevinsen, I think recomputing pointer focus on every surface layout change would be expected. You recompute on every pointer motion as well, right? And motion, when it happens, tends to be much more frequent. So why not?
08:56 pq: kennylevinsen, they can be batched to an output repaint though, so not at protocol rate.
08:56 kennylevinsen: my thought was: If subsurfaces can extend past the window bounds without causing geometry changes and cause immediate pointer focus changes, then you *could* in a highly absurd scenario use transparent subsurfaces to probe cursor position outside the window bounds without the user noticing
08:57 pq: ooh
08:57 kennylevinsen: however, I don't think our justification for the current behavior is anything other than it being our long-standing status quo
08:58 pq: I never thought of that.
09:00 kennylevinsen: We crop things to the toplevel bounds so it's not as much of an issue in sway, but it does give reason for reservation with respect ot the behavior
09:06 pq: It's a good point. I think re-focusing on actual input hardware events is good.
10:01 wlb: wayland-protocols/main: Derek Foreman * various: Fix definition of double-buffered state https://gitlab.freedesktop.org/wayland/wayland-protocols/commit/c5b47dc92810 stable/viewporter/viewporter.xml stable/xdg-shell/xdg-shell.xml staging/tearing-control/tearing-control-v1.xml staging/xwayland-shell/xwayland-shell-v1.xml unstable/pointer-constraints/pointer-constraints-unstable-v1.xml unstable/xdg-shell/xdg-shell-unstab
10:01 wlb: wayland-protocols Merge request !301 merged \o/ (Fix double-buffered state definitions by referring to wl_surface.commit instead https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/301)
10:46 wlb: weston/main: Pekka Paalanen * 16 commits https://gitlab.freedesktop.org/wayland/weston/compare/4193bd660f5e1f1ae61e2a9d2f95afa97d48719c...b0d5b066c585a7a8671663b3db724d1f647588d2
10:46 wlb: weston Merge request !1505 merged \o/ (Program KMS "Colorspace" property from config file https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1505)
16:18 riteo: daniels: all right, thanks
21:13 codecolla: Hi all, my original C code for wayland client was waiting a "wl_shell" in the registry_add_object called by wl_display_roundtrip
21:14 codecolla: That was working at that time but not anymore (project standby for long and now switched to Fedora 40)
21:15 codecolla: this https://bugaevc.gitbooks.io/writing-wayland-clients/content/beyond-the-black-square/xdg-shell.html says "hey, go use xdg, it's unstable!"
21:16 codecolla: Chat GPT is trapped by not being updated, could someone help me?
21:19 codecolla: xdg scares me, and, after all, my only goal is to draw egl scene
21:21 codecolla: Do I need this? how to do?
21:35 bl4ckb0ne: xdg-unstable-v6 is deprecated, you should use xdg-shell instead, its upstream
21:35 bl4ckb0ne: chatgpt is never a good idea for that kind of thing
21:35 bl4ckb0ne: https://github.com/emersion/hello-wayland/blob/opengl-cube/main.c is a good starting point
21:37 codecolla: Yeah, thanks, now looking at cglm/cglm.h dnf provides
21:44 codecolla: No official packages for cglm, do you know a best way to install it on fedora?
21:57 codecolla: Found cglm-devel-0.9.4-1.1.x86_64.rpm libcglm0-0.9.4-1.1.x86_64.rpm on rpmfind, the opensuse Tumbleweed version