Re: Re: mouse cursor problems

On Sun, 15 Feb 2004, Ove Kaaven wrote:

> Well, your patch did not do much to improve this. It still passes a NULL
> HWND to SetCooperativeLevel. And it's not like the current fallback
> SetCooperativeLevel uses when getting a NULL, using GetDesktopWindow(),
> is ever going to be unknown, non-existent, or changed. It's basically a
> constant value. I don't see any problem with using it in alloc_device.

Let me put things in perspective.  When Phil started this thread I
proposed that the problems with the cursors are unnecessary warping.
Later on, I classified the problems into 3 symptoms of which unnecessary
warping was Symptom 3.  Gregory wrote that his game displays Symptom 1
(WineX giving relative data and the game wants absolute data).  I replied
that Gregory's problem is Symptom 3 in disguise.  My patch does nothing
for Gregory other than prove this.  When Symptom 3 is fixed I suspect that
Gregory will not need my patch.
                                                                                
The point of calling SetCooperativeLevel is to get a window handle.
SetCooperativeLevel has a default window handle.  For the time being it is
suitable.  We may wish to change it in future.  This way we will change it
in one place only.
                                                                                
The value of my patch is in pointing out a place where WineX is not up to
spec.  Fixing this deviation will make WineX more robust.  However,
currently my patch does not specifically fix any game that I know of.

The "what's the point?" tone I sense in your letter is somewhat justified.
I admit the point is rather obscure.  I hope Gregory understood.  Knowing
what the problem is, is a big step on the path to fixing it.
                                                                                
Nadav Frum
----------