https://bugs.kde.org/show_bug.cgi?id=520910

Vlad Zahorodnii <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REPORTED                    |CONFIRMED
     Ever confirmed|0                           |1

--- Comment #5 from Vlad Zahorodnii <[email protected]> ---
> heavy shaders/unlimited FPS increases repro rate to 100%.

If it takes longer to render a frame, and the locked pointer is destroyed asap,
I can sort of see why the cursor position is not restored at the center of the
window.

The cursor position hint is synchronized to frames. However, we process the
destruction of locked pointers immediately.

On the other hand, it looks like mutter doesn't implement the cursor position
as a double buffer state property, so it always uses the latest cursor position
hint when the locked pointer is destroyed. Only the constrained region is a
double buffered property.

To be frank, I'm not entirely sure who is right:

- if the cursor position hint is double buffered, I guess kwin can only unlock
the pointer after all frames before zwp_locked_pointer.destroy have been
completed => some latency
- if the cursor position hint is single buffered, it makes things easier with
our current code but it is not quite technically correct

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to