00:24 zamundaaa[m]: Colin Marc: cursors are actually displayed on the screen by the compositor, icons are often shown by other clients
00:25 mclasen: other clients are showing the icons that apps send to the compositor?
00:27 zamundaaa[m]: Yes
00:29 zamundaaa[m]: So icons being only in a format that can be shared to other clients without multiple steps, copies and caveats in between makes much more sense than trying to make surfaces somehow fit something they were never meant for
01:28 whot: pq: gpg key is verified now, htx for the reminder
07:47 pq: colinmarc, cursors are surfaces, because in 2012, surfaces existed, and icons were "invented" just recently. Also in 2012, most of the wl_surface extensions did not exist.
07:52 vyivel: the funniest thing about cursor surfaces is cursor subsurfaces
07:59 kennylevinsen: Funny in an "oh god please stop" kind of way
07:59 vyivel: the only kind of fun around here
08:05 davidre: sync or desync 🤔
08:05 pq: jadahl, could you verify your user ID at keys.openpgp.org? Otherwise they don't distribute your user ID, which means gpg will not import the key.
08:05 jadahl: pq: sure, let me see how to do that
08:06 emersion: they send you an email when you upload your key
08:08 pq: I recently changed my gpg to use trust model tofu+pgp, but even after googling I cannot seem to figure out how to see the tofu statistics at will. I found a tool doing that, written in lisp, calling gpg with a command line argument that does not seem to do anything.
08:09 emersion: lol
08:09 jadahl: pq: the @gmail.com part of the key should be verified now at least
08:10 pq: at best, I've been able to see tofu statistics when I verify a signed git tag :-p
08:10 pq: jadahl, import works, thanks!
08:10 pq: whot, thank you, too!
08:14 pq: the combination of --with-tofu and --with-colons does show more data, but its humanly indecipherable
08:14 pq: because of --with-colons
08:15 pq: maybe the lisp script really is the only way
08:23 jadahl: is --with-tofu some wierd trick to make people hungry?
08:29 daniels: pq: istr the most compelling argument for cursor-as-surface being that you could reuse frame callbacks ofr animated cursors
08:30 pq: sure, that was the only feature that existed
08:32 pq: using a wl_surface makes sense, but we made a wl_window for all the extra stuff that now hangs off wl_surface
08:32 pq: *never made
09:24 colinmarc: thank you all for the answers!
09:27 colinmarc: will cursor shape be implemented/has it been implemented by most frameworks going forward? I'm curious if that will become more standard given that it seems like only games really care about flashy cursors
09:28 d_ed[m]: Cursor shape is implemented in Qt, there is an open MR in GTK
09:29 d_ed[m]: It is also implemented in SDL
09:29 colinmarc: and it'll be the default? or do apps have to opt in?
09:29 d_ed[m]: Default in Qt and SDL when an app asks for a standard cursor
09:29 jadahl: apps can do whatever they prefer, but should absolutely support non-cursor-shape if they have any interest working everywhere
09:30 jadahl: (with everywhere I mean for example LTS distros)
09:32 jadahl: but I imagine defaulting to cursor shapes is what most will do, if thats supported by the compositor
09:33 colinmarc: cool, intuitively it seems much nicer (and it's also much much easier for my specific use case)
09:34 colinmarc: I was just curious since I haven't seen any apps do cursor shape yet (I get the compositor-side implementation for free from smithay)
09:35 davidre: Qt 6.7 does for sure
09:35 davidre: I think it was added with 6.6.
09:39 d_ed[m]: davidre: it was
10:39 wlb: weston Merge request !1534 opened by Lukasz Czechowski (lczechowski) vnc: Allow neatvnc in version 0.8.0 https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1534
10:57 wlb: weston/main: Marius Vlad * gitlab-ci.yml: Update CI_BUILD_REF_NAME https://gitlab.freedesktop.org/wayland/weston/commit/dedcfb40ddfd .gitlab-ci.yml
10:57 wlb: weston Merge request !1529 merged \o/ (gitlab-ci.yml: Update CI_BUILD_REF_NAME https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1529)
11:24 wlb: weston/main: Ray Smith * backend-headless: Don't try to finish frame if it was cancelled https://gitlab.freedesktop.org/wayland/weston/commit/433f648adbff libweston/backend-headless/headless.c
11:24 wlb: weston Merge request !1427 merged \o/ (backend-headless: Don't try to finish frame if it was cancelled https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1427)
11:36 wlb: weston/main: Ray Smith * backend-drm: don't leak gem_handle_refcnt in drm_destroy https://gitlab.freedesktop.org/wayland/weston/commit/80f096d9d207 libweston/backend-drm/drm.c
11:36 wlb: weston/main: Ray Smith * backend-drm: always create gem_handle_refcnt hash table https://gitlab.freedesktop.org/wayland/weston/commit/e79600483ff3 libweston/backend-drm/drm.c
11:36 wlb: weston Merge request !1426 merged \o/ (backend-drm: always create gem_handle_refcnt hash table https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1426)
12:19 wlb: weston Merge request !1535 opened by Michael Tretter (m.tretter) Draft: Scriptable shell with Lua https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1535
12:19 wlb: weston Merge request !1536 opened by Marius Vlad (mvlad) Draft: gitlab-ci.yml: Remove possible duplicate pipelines https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1536 [CI]
12:39 wlb: weston Merge request !1490 merged \o/ (Deprecate weston_layer_entry_insert/weston_layer_entry_remove in favor of weston_view_move_to_layer() https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1490)
12:39 wlb: weston/main: Marius Vlad * libweston/desktop: Use the weston_view_move_to_layer helper https://gitlab.freedesktop.org/wayland/weston/commit/904811100e13 libweston/desktop/xwayland.c
12:39 wlb: weston/main: Marius Vlad * weston-test-desktop-shell: Use the weston_view_move_to_layer helper https://gitlab.freedesktop.org/wayland/weston/commit/d62e8052f000 tests/weston-test-desktop-shell.c
12:39 wlb: weston/main: Marius Vlad * compositor: Remove layer_entry_remove/layer_entry_insert https://gitlab.freedesktop.org/wayland/weston/commit/f2486c8b9623 include/libweston/libweston.h libweston/compositor.c
12:45 wlb: weston Merge request !1340 merged \o/ (kiosk-shell, xwayland: Add Xwayland surfaces appidd based on WM_NAME and WM_CLASS https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1340)
12:45 wlb: weston/main: Marius Vlad * kiosk-shell, xwayland: Add a new weston_xwayland_surface_api function https://gitlab.freedesktop.org/wayland/weston/commit/619578f62f96 include/libweston/xwayland-api.h kiosk-shell/kiosk-shell.c kiosk-shell/kiosk-shell.h xwayland/window-manager.c
12:45 wlb: weston/main: Marius Vlad * kiosk-shell: Check both WM_NAME and WM_CLASS at the same time https://gitlab.freedesktop.org/wayland/weston/commit/bdcbdb65200f kiosk-shell/kiosk-shell.c
12:45 wlb: weston/main: Marius Vlad * doc/sphinx/toc/kiosk-shell.rst: Add doc for Xwayland appidis https://gitlab.freedesktop.org/wayland/weston/commit/e80ebde0285a doc/sphinx/toc/kiosk-shell.rst
13:08 wlb: weston/main: Lukasz Czechowski * vnc: Allow neatvnc in version 0.8.0 https://gitlab.freedesktop.org/wayland/weston/commit/05e540565105 libweston/backend-vnc/meson.build
13:08 wlb: weston Merge request !1534 merged \o/ (vnc: Allow neatvnc in version 0.8.0 https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1534)
15:36 wlb: wayland-protocols Issue #201 opened by Simon Ser (emersion) xdg-toplevel-icon: buffer release semantics are undefined https://gitlab.freedesktop.org/wayland/wayland-protocols/-/issues/201
16:36 wlb: weston Merge request !1537 opened by Daniel Martinez (Calandracas) Fix Surface Viewport Memory Leak https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1537
17:01 wlb: weston Merge request !1538 opened by Derek Foreman (derekf) Draft: libweston: remove weston_surface_copy_content https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1538 [GL renderer], [IVI shell], [libweston API], [Pixman renderer]
19:23 Vaughn: I ran into a proton/xwayland issue with nvidia 555 that I think is probably a bug in Proton.
19:24 Vaughn: So, not really your problem. But since I'm both curious and not familiar with graphics programming at all, I'm hoping someone equally curious would like to give my diagnosis in https://github.com/ValveSoftware/Proton/issues/7580#issuecomment-2155317373 a glance.
19:24 Vaughn: It _seems_ simple enough, on the face of it.
19:33 zamundaaa[m]: Vulkan fences don't really have anything to do with explicit sync with the window system
19:35 zamundaaa[m]: Without looking much further into the bug, I agree though that the check for success is wrong
19:36 zamundaaa[m]: eh, no, it checks for the result of the UNIX_CALL, not for the Vulkan function. Might be a bug somewhere else in Proton
20:45 Vaughn: Too many layers of redirection. Still, it's definitely that assert that's failing. It seems fair to assume that it's supposed to be returning the actual return value of vkWaitForFences...
20:46 Vaughn: Someone on the Proton team will need to take a look.
20:47 Vaughn: Hmm, no, wait. The result is returned out-of-band.
21:39 JoshuaAshton: Vaughn: It just means a segfault happened on the other side.
21:39 Vaughn: I see.
21:39 Vaughn: Sure would've been useful if gdb had worked. :/
21:40 JoshuaAshton: you need to do some incantation to make that work with wine iirc
21:47 Vaughn: The documentation currently claims it 'should just work'. :-)
21:47 Vaughn: Regardless, I tried the listed steps for 8.x (on 8.x) and failed. I'm used to that sort of thing happening because I'm on nixos, so I didn't dig too far.
21:50 Vaughn: Thanks for the insight, in any case. I don't think I can do much with this, but it's still nice to know.
23:48 wlb: wayland Issue #470 opened by Dennis Lonoshchuk (laycookie) Identifying which wl_output has the pointer. https://gitlab.freedesktop.org/wayland/wayland/-/issues/470