Event trail - better patch

[Brad Campbell]

Ove Kaaven wrote:
> 
> On Wed, 29 Aug 2001, Brad Campbell wrote:
> 
> > Ove Kaaven wrote:
> > >
> > > On Wed, 29 Aug 2001, Brad Campbell wrote:
> > >
> > > > Ok, I'm now at a dead loss.. Where does the message sent by X11DRV_SendEvent go?
> > >
> > > Through a lot of indirection, it is dispatched to mouse_event() in
> > > windows/input.c.
> >
> > Thankyou.. I'm looking at it now..
> >
> > This patch solves the problem without sending extra messages.. would this
> > be a resonably correct fix and shuold I post it to the developers list?
> 
> Would work for the x11drv in normal operation, where MOUSEEVENTF_ABSOLUTE
> is always set, but not in situations where it's not set (like DGA
> perhaps?). Maybe it's best to force a movement message after all? Hmm,
> does Real Windoze force one before the button down message is received? If
> not, you may have to do the movement processing (only done for the
> MOUSEEVENTF_MOVE right now) for all events, not just the movement. You
> could ask on wine-devel.

I have done just that, thanks..

I ran a trace on a test program I knocked up, using winsight32 from Borland
Delphi4, and did not spot a motion event before the button press, but then
this is not as detailed a trace as I can get with +event,+message,+cursor and
some additional traces I added in for good measure.
Sould it not be nice to be able to do that to windows :p)
I'll try the approach of running the movement processing for all events now.

Well, if nothing else, I have learend an awfull lot about where to look and
what traces to run to figure out where to look in the code.


-- 
Brad....
                   /"\
Save the Forests   \ /     ASCII RIBBON CAMPAIGN
Burn a Greenie.     X      AGAINST HTML MAIL
                   / \