[19.09] nixos/libinput: apply options to all device types

[worldofpeace]

Motivation for this change

Without this change any special config used for libinput like accelSpeed won't apply to mice because it isn't a touchpad.

This could possibly change the user experience for someone, like if they came to rely on the bug.
An example situation would be:

• have services.xserver.libinput.accelSpeed in config
• not notice it doesn't apply to your device that isn't a touchpad (mouse)
• update on 19.09 and suddenly the acceleration of your mouse is different.

It's generally not good to make a change like that, but it would be nice to be able to actually configure your mouse with libinput...

Things done

• Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS linux)
• Built on platform(s)
  • NixOS
  • macOS
  • other Linux distributions
• Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
• Tested compilation of all pkgs that depend on this change using nix-shell -p nix-review --run "nix-review wip"
• Tested execution of all binary files (usually in ./result/bin/)
• Determined the impact on package closure size (by running nix path-info -S before and after)
• Ensured that relevant documentation is up to date
• Fits CONTRIBUTING.md.

Notify maintainers

@NixOS/backports (and myself apparently)

[lheckemann]

This stops libinput from being set as the driver for all touchpads, so I think we'd need to split that part into a new InputClass section.

I don't think this should go to 19.09 though, since as you said it changes existing behaviour and configuring mice using libinput is possible on 19.09 using services.xserver.inputClassSections.

[worldofpeace]

This stops libinput from being set as the driver for all touchpads, so I think we'd need to split that part into a new InputClass section.

Regression ❔ (I'd definitely not want to backport that)

[lheckemann]

Yes, and that should be fixed on master as well (has it already been merged to master?).

[worldofpeace]

Yes, and that should be fixed on master as well (has it already been merged to master?).

If you look at the commit is says it has been cherry-picked.

Though what is confusing to me is that I tested the commit functioning, and llibinput list-devices still lists my touchpad etc.

[worldofpeace]

We've established there was no issues at #72774 (comment).

[thiagokokada]

Just to leave it here since this change surprised me. I was actually reliant of the fact that those options are only applied to a touchpad, since I want to enable natural scrolling in my touchpad, however not in my mouse. It was really difficult to understand what was happening.

While it maybe just me, maybe there is more people that had this same issue.

[lheckemann]

This is exactly why I didn't want it backported…

[thiagokokada]

Well, actually my setup is broken. I can't set natural scrolling to work only in my touchpad and not in mouse. It is either all devices (using services.xserver.libinput.naturalScrolling = true) or nothing. I tried both services.xserver.extraConfig and services.xserver.inputClassSection and they both don't work.

Could we revert this PR @lheckemann @worldofpeace? Or at least, there is some solution for this problem? It is really awful to have natural scrolling for both mouse and touchpad, however touchpad without natural scrolling is very strange.

[thiagokokada]

BTW, I have a hypothesis why this is either all or nothing. It is because NaturalScrolling option for libinput is always set, so even if you don't set services.xserver.libinput.naturalScrolling it will be set to false, applying this setting to all devices.