Repository navigation
Conversation
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
A terminal that maps again as it closes lands on special:screensaver, which the class rule keeps out of sight. Focusing it there showed that workspace, and on a monitor with the scratchpad open it took the scratchpad's place and left it hidden after the screensaver was gone. Reproduced on a worker on Hyprland 0.56.2 by mapping a screensaver-class window with the scratchpad showing; with this guard the scratchpad stays. Co-Authored-By: Codex Medium <noreply@openai.com> Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Reviewed and run on a disposable Omarchy worker (Hyprland 0.56.2, foot, one 1280x800 output) with Omarchy pointed at this branch. Reproduced. With the network panel or the calendar popup open, One fix pushed (4c37417): the handler now skips a screensaver that maps on Tests on the worker: Second opinion: Codex Medium reviewed it in two rounds. In the first it found the hidden-workspace problem above, which the worker run then confirmed. Its second round, on 4c37417, found nothing open. Codex can read this session's work, so its agreement is not guaranteed to be independent. Related: this now closes #13265 and #8291, whose calendar case was reproduced and seen fixed. #6917 has the same mechanism but was not run with the power panel, so it is not linked. #13475 also targets #13265 by closing open panels before launch and giving the screensaver a five-second focus grace. Codex judged this pull request the more direct fix for the fullscreen problem. #11239 (closing panels drawn over the screensaver) and the focus-history fixes for the exclusive-focus menu case (#7102, #13637, #12258) address other parts and compose with this one. Waiting on the maintainer: the change is verified, and the failing |
|
|
This pull request no longer has anything to do on Since it was reviewed, fdbe182 ("Launch the native Wayland screensaver") replaced the terminal screensaver with a native Wayland overlay. Checked by reading What this does not settle: whether #6917 and #13265 still happen with the native overlay while a bar panel holds keyboard focus. The terminal mechanism behind them is gone, but the overlay lives in the Thanks @seantimm for the careful measurements in this one. Waiting on the maintainer, with the recommendation that it be closed. |
Fixes the fullscreen half of #6917 (also reported as #13265 and #8291).
Why
When a layer surface holds keyboard focus (an open bar panel, the Omarchy menu), Hyprland maps the screensaver unfocused and skips its
fullscreenwindow rule. The screensaver shows as a small floating window at the terminal's default size, 700x500 for foot.The open PRs for #6917 deal with the screensaver quitting after a second (#7102, #12258, #13637, #13475) or with the panel staying on screen (#11239). None of them makes the window fullscreen, and the thread on #6917 treats that part as out of Omarchy's reach. It is reachable: the window rule is skipped, but an explicit fullscreen dispatch for the same window is honoured.
How
default/hypr/apps/system.luagains awindow.openhandler next to the screensaver window rules. For a window of classorg.omarchy.screensaverit dispatches focus and then fullscreen, both addressed to that window.Testing
Measured on Hyprland 0.56.2, foot, one 5120x2160 output at scale 2, running
bin/omarchy-launch-screensaver forcefrom this branch's base (8b4eae66), no input given. Window state sampled every 250 ms fromhyprctl clients -jandhyprctl activewindow -j.quattroomarchy-keyboard-panel)fullscreen: 0, never active, gone by 1.5 sfullscreen: 2, active, still up at 4 somarchy-menu)fullscreen: 0, never active, gone by 1.5 sfullscreen: 2, never active, gone by 1.5 sfullscreen: 2, active, still up at 4 sWith the network panel open, a Space keypress injected through
ydotooldismissed the screensaver, so keys reach its terminal once it has focus.test/shell.d/screensaver-fullscreen-test.shloadssystem.luaagainst a stubhland checks that an opening screensaver gets both dispatches, addressed to itself, and that another window gets none. It passes, as doscreensaver-workspace-test.sh,hyprland-default-config-test.shandhyprland-window-test.sh../test/allfails in 11 of 287 test files here, none of them about Hyprland config or the screensaver:config,elsewhen-migration,kernel-headers-migration,launch-browser,network-captive-portal,omarchy-kernel-migration,passwordless-grant-lifecycle,screenshot-sanity,snapper,unowned-system-paths,video-background. I did not rerun the suite on a cleanquattrocheckout to compare.Known limits
🤖 Generated with Claude Code
Fixes #13265
Fixes #8291