Gerard,
Here's some comments on my last patch:
We have to use SendMessage instead of SendNotifyMessage for now, because
SendNotifyMessage is not implemented correctly (it uses PostMessage). That
was a mistake in my previous patch.
Also, we should send WM_SYNCPAINT only when event->count is zero. That
should help the performance.
Then we should do the modification also in: EVENT_GraphicExpose() function.
Stephane Lussier.
Macadamian Technologies
> -----Original Message-----
> From: gerard patel [mailto:[EMAIL PROTECTED]]
> Sent: Sunday, July 23, 2000 10:06 AM
> To: Stephane Lussier
> Cc: [EMAIL PROTECTED]
> Subject: RE: EventExpose race condition problem
>
>
> At 03:06 PM 7/17/00 -0400, you wrote:
> >Ian,
> >
> >This patch should solve the splash screen problem you're
> experiencing. Let
> >me know if it works.
> >
>
> Next episode of the EventExpose saga...
>
> I've done more testing with my Wine crasher (btw if someone wants
> it, I can
> send it, it's a chess program from the example code of Borland C 3.1).
> When it starts, it outputs a message box (opening book not
> found). If I move
> this messagebox window (by holding it in its caption), it leaves
> often fine marks
> on the main window, especially in the menu area.
> If I move the message box very fast, sometimes Wine goes in real
> trouble (repainting
> does not occur anymore). This is leading me to think that there is some
> reentry problem somewhere.
>
> I've tried a few changes, with no great success. The best result
> I had was by
> replacing SendNotifyMessage by SendMessage. It fixes my original
> repainting
> problem without creating too much elsewhere (with all my tests I
> had 2 small
> problems, one time Notepad 16 bits did not refresh itself after
> switching from
> another task, another (with Acrobat viewer 32) I had an error
> message about
> an invalid region. I can't be sure that these problems are
> related to all these
> changes, though; and I can't reproduce them :-/)
>
> Gerard
>
