Re: [hatari-devel] Mouse pointer sync (was: Strange timestamp issue)

[Thomas Huth <th.huth@xxxxxxxxx>]

Am Fri, 2 Aug 2024 10:29:38 +0200
schrieb Uwe Seimet <Uwe.Seimet@xxxxxxxxx>:

> > > This sounds like an interesting idea. And indeed, the behavior of the mouse
> > > pointer is better now. Without the patch after moving the mouse pointer out
> > > of the Hatari window and moving it around on the Linux desktop, the pointer
> > > was immediately out of sync after re-entering the Hatari window. Now as long
> > > as no other windows are involved it is much better. But as soon as I open
> > > another window that covers the Hatari window, everything is out of sync
> > > again. Looks to me as if some movements get lost in this case?  
> > 
> > I tried a couple of times, but I cannot reproduce that problem with the
> > covering window here (using Linux with XFCE). Maybe it depends on the
> > desktop environment?
> > 
> > So no clue how to fix that problem... maybe you could try whether it works
> > any better if you add the new code only to SDL_WINDOWEVENT_LEAVE and
> > SDL_WINDOWEVENT_ENTER instead of executing it for SDL_WINDOWEVENT_FOCUS_GAINED
> > and SDL_WINDOWEVENT_FOCUS_LOST, too? ... I doubt that such a change will
> > help, but it is maybe still worth a try.  
> 
> This change helps, indeed. The problem caused by windows on top of the
> Hatari window is gone. If I re-add SDL_WINDOWEVENT_FOCUS_GAINED and
> SDL_WINDOWEVENT_FOCUS_LOST it's there again. (I'm using KDE.)
> All in all, with these changes my initial mouse pointer problems appears to
> be resolved, which is great!

Sounds good! And the fix was easier than expected :-)

> Please tackle the fast-forwarding problem/timestamp problem next ;-). If
> it's not too much work it would be nice to test whether things improve when
> the behavior of the system timer is changed, as Thorsten suggested.

I've got other things on my todo list first, but maybe Thorsten could
provide a patch if he has some ideas how to implement this.

 Thomas