hyprland 0.52.0-1 e uwsm #12244
Replies: 4 comments 7 replies
-
Beta Was this translation helpful? Give feedback.
-
|
And then it's time to use Rofi. Ok got it :) |
Beta Was this translation helpful? Give feedback.
-
|
Just tested this on 0.52.0, with pcmanfm and firefox, started via a keybind, with both Can you post |
Beta Was this translation helpful? Give feedback.
-
|
I found a temporary workaround for launching apps under uwsm from Hyprland keybinds. bind = $mainMod, T, exec, rofi -modi run -show run -filter "uwsm app -- $terminal" Then pressing Enter launches the app under its own uwsm systemd scope. It’s not automatic (requires pressing Enter once), but it ensures the app is spawned correctly under uwsm instead of Hyprland’s process tree. It also works with App2Unit instead of uwsm app -- |
Beta Was this translation helpful? Give feedback.

Uh oh!
There was an error while loading. Please reload this page.
-
Description
Since version 0.52.0-1, Hyprland seems to execute exec commands differently.
When I launch applications through Hyprland binds using uwsm app --,
the apps are started as children of the Hyprland process instead of being placed in their own
app-*.scope under systemd.
However, if I run the same uwsm app -- command from a terminal that was launched via Rofi,
the app is correctly started in a new scope.
If I run the command from a terminal that was itself launched via a Hyprland bind,
then the new app also stays under the Hyprland cgroup — not under systemd.
So this behavior appears to propagate through the process tree originating from Hyprland itself.
Steps to Reproduce
Add this to ~/.config/hypr/hyprland.conf:
Press SUPER+Return to open Kitty.
Inside Kitty, run:
Check with:
The app runs under Hyprland’s cgroup, not its own app-*.scope.
Now, run the same uwsm app -- firefox from a terminal opened via Rofi:
it correctly spawns a new app-*.scope.
Expected behavior
Apps launched through uwsm app -- should always get their own systemd scope,
regardless of whether they originate from a Hyprland bind or a terminal opened via Hyprland.
Actual behavior
Anything launched directly or indirectly from a Hyprland bind stays in Hyprland’s cgroup.
Versions
Hyprland: 0.52.0-1 (Arch Linux)
uwsm: [exact version]
systemd: [exact version]
wlroots: [optional]
Distribution: Arch Linux (fully up to date)
Additional notes
This behavior did not occur with Hyprland 0.51.x.
It seems that exec-spawned processes now inherit Hyprland’s session leader
instead of being re-parented under systemd’s app scopes.
Beta Was this translation helpful? Give feedback.
All reactions