Re: dlls/x11drv/mouse.c

[Alexandre Julliard]

Ulrich Czekalla  writes:
> We had a XFlush call in X11DRV_SetCursorPos before but it was removed for
> some reason. This restores the call and fixes a regression in JInitiator
> where the mouse position was being reported incorrectly.

This sounds like it's just masking the problem. Do you know precisely
why the mouse position is incorrect?

dlls/x11drv/mouse.c

[Andreas Mohr]

Hi,

On Fri, Sep 09, 2005 at 11:22:26AM +0200, Alexandre Julliard wrote:
> Ulrich Czekalla  writes:
> > We had a XFlush call in X11DRV_SetCursorPos before but it was removed for
> > some reason. This restores the call and fixes a regression in JInitiator
> > where the mouse position was being reported incorrectly.
>
> This sounds like it's just masking the problem. Do you know precisely
> why the mouse position is incorrect?

In case further testing/evidence is needed,
http://www.dicker-turm.de/panorama-dicker-turm.exe
has mouse position problems on Wine as well (Windows is ok) which could
be the same issue.

Andreas

dlls/x11drv/mouse.c

[Ulrich Czekalla]

On Fri, Sep 09, 2005 at 11:22:26AM +0200, Alexandre Julliard wrote:
> This sounds like it's just masking the problem. Do you know precisely
> why the mouse position is incorrect?

No :-/ I can't explain it. The XQueryPointer call in X11DRV_GetCursorPos is
returning an incorrect value. I don't see anything wrong with the way we
are calling it. Maybe it's a bug in X.

Older versions of Wine worked fine so I looked at what has changed since it
broke and concluded that the removal of the XFlush call provokes the
problem.

/Ulrich

dlls/x11drv/mouse.c

[Alexandre Julliard]

Ulrich Czekalla  writes:
> No :-/ I can't explain it. The XQueryPointer call in X11DRV_GetCursorPos is
> returning an incorrect value. I don't see anything wrong with the way we
> are calling it. Maybe it's a bug in X.

Would you be able to write a small test case demonstrating the problem?