STEPS TO REPRODUCE 1. With a touchscreen laptop, configure a multi-head setup with the laptop on the right and an external (non-touch) monitor on the left. 2. Try to use the laptop's touchscren to control widgets located there. OBSERVED RESULT The touch location is interpreted as if the touchscreen covered the entire width of the virtual screen with two displays. For example, a touch in the middle of the laptop's touchscreen occurs (from Plasma's point of view) on the borderline between the two displays, i.e., in the middle of the entire large virtual screen. EXPECTED RESULT Touches should be interpreted only within the touchscreen that generates them; a single touchscreen's control input should never be stretched to cover an entire virtual screen. SOFTWARE/OS VERSIONS Linux: 4.19.9 (ArchLinux) KDE Plasma Version: 5.14.4 KDE Frameworks Version: 5.53.0 Qt Version: 5.12.0
Workaround: you need to issue an xinput --map-to-output command after changing the arrangement of the displays on X11. KDE should do this automatically. (Wayland seems to get it right.)
I can confirm that the problem is still there on plasma 5.20.3 on X11 (works on wayland) The workaround is working for me, thanks (on a Dell 7410 laptop): xinput --map-to-output "Wacom HID 4957 Finger touch" eDP-1 I tested it on a fedora 33 system; also tried with the plasma-unstable repo to use the latest version of plasma (copr mkyra)
*** This bug has been marked as a duplicate of bug 429714 ***