Mihai Fufezan

I believe it's best to reuse the nix to hyprlang translator that the hyprland module uses. Perhaps externalize it to a separate file/lib function. That way, we won't have to...

Imo we shouldn't break backwards-compatibility. I managed to fix this in a different way. ```diff diff --git a/modules/services/window-managers/hyprland.nix b/modules/services/window-managers/hyprland.nix index 168bab65..6bd56fd0 100644 --- a/modules/services/window-managers/hyprland.nix +++ b/modules/services/window-managers/hyprland.nix @@ -211,20 +211,32 @@...

I left the patch here in case @matt1432 wanted to update the PR to use some ideas from it (or use it entirely).

```nix device = [ { name = "mouse1234"; option1 = "value1"; } { name = "keyboard_something"; kb_layout = "colemak"; } ]; ``` Just like any other duplicate key.

Really? Didn't know that. @rycee would you know how to work around this limitation, or have I messed up somewhere?

Does that work in case `device` isn't a list?

That's great to hear! However, it won't completely replace the NixOS module. It still enables `hardware.opengl`, `security.polkit` and makes the sessionPackage visible to display managers.

That's a good idea. We could raise this issue with ReGreet/greetd, but I'm not sure about other display managers. The issue is probably irrelevant outside of immutable distros.

> Probably makes this line unnecessary: If it's the case, can you also remove it in this PR?

> * Does/will the EWW version continue to receive updates? Probably not, as I barely have time to work on my ags config. > * Do you intend to upstream...