Bug 425758 - Edge detection pushback based on wrong monitor in overlapping setups
Summary: Edge detection pushback based on wrong monitor in overlapping setups
Status: RESOLVED FIXED
Alias: None
Product: kwin
Classification: Plasma
Component: multi-screen (show other bugs)
Version: 5.18.5
Platform: Gentoo Packages Linux
: NOR normal
Target Milestone: ---
Assignee: KWin default assignee
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2020-08-24 20:25 UTC by Ian Silvester
Modified: 2024-06-07 17:00 UTC (History)
4 users (show)

See Also:
Latest Commit:
Version Fixed In:


Attachments
Video demonstrating the bug (56.11 KB, video/x-matroska)
2020-08-24 20:25 UTC, Ian Silvester
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Ian Silvester 2020-08-24 20:25:09 UTC
Created attachment 131157 [details]
Video demonstrating the bug

SUMMARY
Cursor jump when raising auto-hiding taskbar panel when using external monitor as primary display

STEPS TO REPRODUCE
1. Attach an external monitor, configure it as enabled and primary, disable laptop screen.
2. Configure taskbar panel to auto-hide.
3. Move cursor slowly towards bottom of screen.


OBSERVED RESULT
When bottom edge of screen is reached, the taskbar panel will show but the cursor will jump around one inch up the screen so it is no longer floats over the panel, meaning it then auto-hides once more.

EXPECTED RESULT
When bottom edge of screen is reached, the taskbar panel will show and the cursor will not move.

SOFTWARE/OS VERSIONS
Linux/KDE Plasma: 4.19.52
(available in About System)
KDE Plasma Version: 5.18.5
KDE Frameworks Version: 5.70.0
Qt Version: 5.14.2

ADDITIONAL INFORMATION
See attached video for a demonstration.
Comment 1 Nate Graham 2021-08-17 02:05:37 UTC
But only with an external screen, right? Any chance you can try again in Plasma 5.22? Does it happen in the Wayland session?
Comment 2 Ian Silvester 2021-08-17 04:04:30 UTC
Hi Nate,

The bug is unchanged in Plasma 5.22 unfortunately, and yes only occurs when an external monitor is the primary display.

I don't know whether it affects the Wayland session. I'll have to read up on how to try that out instead of X then follow up.

Ian
Comment 3 Nate Graham 2021-08-17 15:20:03 UTC
Gotcha, thanks for the info. I would be interested to know if it happens on Wayland as well.
Comment 4 Ian Silvester 2021-08-22 17:39:35 UTC
Thanks Nate,

I've opened a forum post (https://forum.kde.org/viewtopic.php?f=111&t=172386) to get some assistance in running Plasma with a Wayland session (do you have any ideas?). It'll be useful to get it working because then I can help with the push towards Wayland in general!

Will follow up when I've got Wayland working,

Ian
Comment 5 Ian Silvester 2021-08-22 19:27:51 UTC
Typically, just after I post I get things working :-D

I now have Wayland working under (I believe) XWayland - that is to say within my Plasma X session I have a window that contains an entire Plasma Wayland session. So long as that is a fair test, I can tell you that the bug does also affect the Wayland session :-/

Hope this helps,

Ian
Comment 6 Ian Silvester 2021-08-24 15:23:00 UTC
Thanks for the insight as to what might be causing this behaviour David (and consequently renaming the bug). I did some further testing to explore this line of thought.

tl;dr I have good news for the KDE project but bad news for me (and perhaps others who use Lenovo USB C external docks). Plugging the external monitor direct to the HDMI port on my laptop I could not reproduce the behaviour. Connecting via DisplayPort via the Lenovo USB C external dock I can reproduce the behaviour. Ergo it is either an artifact of DisplayPort connections (perhaps someone else can test this directly whose laptop has the necessary hardware) or else it is a weird fault related to forcing video signals via USB C.

For full disclosure, none of my display overlapped one another in my display configuration, and the laptop screen being disabled or enabled did not play into the behaviour.

I further noted that the behaviour is only reproduced when using an external pointing device (also connected via the dock) and is related to the cursor speed (or perhaps acceleration). If I approach the bottom of the primary monitor extremely slowly the behaviour does not occur - it is only when the cursor is moving with some speed that the behaviour occurs. 

If you feel there's enough information to close this bug as 'Could not reproduce' (and I suggest rename it once again!) I will completely understand :o)
Comment 7 Zamundaaa 2024-06-07 17:00:40 UTC
Screen edge behavior with multi monitor setups has been changed significantly in Plasma 6.1, and I'm relatively sure this is fixed now. Just reopen this if you can still reproduce the issue