Created attachment 178770 [details]
Video of cursor jumping when hovering taskbar with 0 padding, note mouse is only ever moved vertically.

SUMMARY
I've noticed my mouse cursor suddenly being displaced to the left occasionally when hovering the task bar, which is quite annoying.  However, this did not happen always, so it took a while to narrow it down.  I've now found a way to reproduce this relatively reliably -- still can take a couple of attempts to trigger the effect; not sure if this is based on timing or some specific cursor positioning/movement.

STEPS TO REPRODUCE
1. Configure taskbar to: position = bottom, alignment = center, width = fill width, visibility = auto hide (!), opacity = opaque, style = floating.  I have not tested all permutations, but I'm pretty sure only bottom and especially auto hide matters here, and I checked and at least style does not matter (also happens when not floating).
2. Dock any window to at least the bottom and top edge of the screen with zero padding.  So either drag the top and bottom window edges to the top and bottom of the screen, or use tiling areas with padding of the tiling zones set to zero.  Note that this does *not* happen when using tiling zones and padding is non-zero.  Note this also does *not* happen if window is only docked to bottom of screen.
3. Move mouse cursor vertically from window to task bar / very bottom of screen to trigger task bar becoming visible, if necessary repeatedly until mouse jump to the left is observed.  Gut feeling happens 2/10 times on average.

OBSERVED RESULT
Mouse cursor sometimes repositions to the left (still at bottom edge of screen) even if mouse is only moved vertically/up and down.

EXPECTED RESULT
Mouse cursor does not reposition to the left if mouse is only moved up and down, but only moves according to actual mouse movement.


SOFTWARE/OS VERSIONS
Operating System: Manjaro Linux 
KDE Plasma Version: 6.2.5
KDE Frameworks Version: 6.10.0
Qt Version: 6.8.2
Kernel Version: 6.12.12-2-MANJARO (64-bit)
Graphics Platform: Wayland

ADDITIONAL INFORMATION
The attached video showcases how this does not happen when padding for tiling zones is non-zero, then the padding is reconfigured to zero and the effect is demonstrated (and also that it can take a few attempts to get it to trigger).  For brevity there is a limited number of demonstrations for the non-zero padding case.  I've tried *a lot* and it didn't trigger in the non-zero case even after several dozen attempts.

I'm not entirely sure which component this relates to specifically, so I'm hoping this can be moved if necessary.