X.Org Server 1.19.1 For Testing On Arch, Libinput X.Org Driver To Be Used By Default

[GreenByte]

libinput is only supported on wayland, the KCM will only work on wayland, not for the xinput driver. xf86-input-libinput is not supported by Qt. So just forget about that on xorg and wait for wayland.

Really? I have been using libinput instead of synaptics and evdev on arch for over 2 years without problems. All the trackpad sections seems to work on mate and gnome shell. I am not sure about kde, because all the previous times I tried it were awful.

[GreenByte]

how else are you going to calibrate a Joystick?

Also when it comes to mouse handling, libinput sucks ass. Who in their right mind would enable mouse acceleration BY DEFAULT and then make it nearly impossible to switch off?

Wow, I have never noticed any mouse acceleration and I have never changed the configuration.

[Guest]

a little offtopic:
does anyone know if mutter is going to support bypassing on Xwayland soon?

I am guessing you are referring to this bug (just register and CC yourself to following updates) https://bugzilla.gnome.org/show_bug.cgi?id=759538
Looking at the first comment though suggests it may not happen soon. Apart from that, this won't affect native wayland applications which won't be be using the xwayland process.

[davidbepo]

I am guessing you are referring to this bug (just register and CC yourself to following updates) https://bugzilla.gnome.org/show_bug.cgi?id=759538
Looking at the first comment though suggests it may not happen soon. Apart from that, this won't affect native wayland applications which won't be be using the xwayland process.

no, i dont mean that, i mean compositor bypassing that eliminates fullscreen overhead
thanks anyway

[Guest]

Ah. Sorry, no idea then. Maybe ask mutter developers on irc.gnome.org/#gnome-shell

[GrayShade]

how else are you going to calibrate a Joystick?

Also when it comes to mouse handling, libinput sucks ass. Who in their right mind would enable mouse acceleration BY DEFAULT and then make it nearly impossible to switch off?

Umm.. Windows?

libinput supposedly recently got much better at touchpads recently. I've been using it with mice for two years or so with no issues (though I don't play games), but I found it terrible with touchpads. I don't mind having acceleration, but it felt "reversed". Even when trying to move the finger very slowly I used to overshoot. It was like the slower I moved my finger, the faster the mouse cursor got.

[ⲣⲂaggins]

libinput is only supported on wayland, the KCM will only work on wayland, not for the xinput driver. xf86-input-libinput is not supported by Qt.

By 'support' I take it you mean 'ability to configure', not 'use as input driver'?

I'm also having a hard time believing what you wrote, as the workaround with xinput set-prop wouldn't make sense in that case.

[tegs]

@Detructor,

Using libinput for mouse is a HUGE improvement than what was there before. You do know you can easily disable mouse acceleration right? Do a Google search for Arch mouse acceleration to bring up their wiki.

[RealNC]

I'll take libinput over synaptics when it's ready, but forcing the change through now seems a little premature if you're using kde. Last time I checked, system settings crashes whenever you enter the touchpad section because of a missing kcm. Since the defaults are awful, this makes it very annoying to use. Even if you track down and compile the kcm, you find that it's incomplete anyway:



Can't be very difficult to fix. Maybe this change will motivate the libinput kcm maintainers to finish it.

You can't even configure the mouse in KDE. The touchpad is the lesser problem here.

Don't like the "single-click acts as double-click" thing KDE uses by default? Too bad. You can't disable that through the GUI.

[ⲣⲂaggins]

Don't like the "single-click acts as double-click" thing KDE uses by default? Too bad. You can't disable that through the GUI.

What's this?