On 1/31/23 7:28 AM, E. Liddell via tde-users wrote:

Maybe you're somehow triggering a keyboard shortcut? (Sounds weird, I know, but they're all handled by libinput these days . . .)

A good start, thank you!

Comically, to look for the related shortcuts I twirled my office chair to the laptop where I am testing TDE. As soon as I touched the mouse -- boom, smacked into desktop 2.

I removed all of the related Ctrl+Fxx keyboard shortcuts. I'll see how that goes. If the behavior stops then I have to figure out what is triggering the shortcut. Trackpad is disabled....

On 2/1/23 1:07 AM, Dr. Nikolaus Klepp via tde-users wrote:

Maybe you need to disable powersave modes for USB in the BIOS. I had the jumping-cursor-problem, too, and it got worse with newer kernel versions - from 6.0 on I got a new keyboard-dies-unexpectedly "died" unexpectidly-problem, turned out it went to sleep and never woke up. Since I disabled USBP powersave the problems are gone.

Thanks! A sleeping USB mouse might be the cause although I never before witnessed this behavior.

In my first attempt to resolve the problem I renamed two xorg.conf.d files that are installed by the libinput package, which basically disables libinput. On the problematic distro the problem did not occur last night, but this likely will require several sittings of observation.

The mention of libinput in the previous reply got me thinking about the basic behavior. My primary TDE test machine is multi-boot. I can boot into an older distro that has the same version of TDE but does not use libinput. I only observed last night so insufficient data, but with the older distro not once did I experience the spastic pointer movement or dumping me into desktop 2.

Thus far I only witness the behavior in TDE and no other DE. Yet the previous reply got me thinking that the behavior was very much like a gesture. To my knowledge nothing on the system, including TDE, has any kind of gesturing enabled. Yet that is what the behavior seems like. Seems libinput is a prime suspect.

From what I have read libinput is not fully developed yet and is useful mostly with laptops and trackpads. I disable trackpads whenever possible and use an external mouse. Disabling libinput is not step backwards for me.

I will tinker with USB power saving some time later because I want to avoid introducing two concurrent troubleshooting variables. I can envision where the power saving might act similar to a gesture. If the problem continues then libinput is not the cause and on to testing USB power saving. Likely I will test by restoring libinput and disabling the power saving and see what happens.

Perhaps there is some kind of interaction between USB power saving and libinput.