rope/nixos
1
1
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

hyprland: drop cursor flips that didn't move the needle

Browse source 
no_hardware_cursors and enable_hyprcursor were chased while debugging
input latency; neither made a perceptible difference. Falling back to
defaults keeps the config honest about what's load-bearing. vrr=2 and
render.direct_scanout stay — direct_scanout is a real fullscreen win
and vrr=2 is a harmless no-op until a driver update lets it engage.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
ediblerope 2026-05-11 10:46:48 +01:00
parent 3c9678ea74
commit 4754c5ceb0
1 changed files with 0 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

9
settings/hyprland.nix
View file

@ -146,15 +146,6 @@
vfr = false; # keep compositor ticking, don't idle between frames vfr = false; # keep compositor ticking, don't idle between frames
}; };
cursor = {
# Counterintuitive on AMD: HW cursor plane can micro-stutter from
# atomic KMS commits, while SW cursors get redrawn at the
# compositor's 190 Hz rate. Many AMD users report SW feels
# smoother. Flip back to 0 if you'd rather try the HW path.
no_hardware_cursors = true;
enable_hyprcursor = false;
};
# Mirror the GNOME bindings so muscle memory carries over # Mirror the GNOME bindings so muscle memory carries over
bind = [ bind = [
"$mod, T, exec, $term" "$mod, T, exec, $term"