From: mentalg@geocities.com
  To  : ggi-develop@eskimo.com
  Date: Tue, 15 Sep 1998 18:21:41 -0400 (EDT)

Re: SYNC mode should go

On Tue, 15 Sep 1998, Marcus Sundberg wrote:

> becka@rz.uni-duesseldorf.de wrote:
> > 
> > Hi !
> > 
> > > > > But the only correct solution is for applications that doesn't
> > > > > continously update the framebuffer to listen for the ggi_expose_event,
> > > > > which we should make the X-targets emit properly.
> > > > Okay, good.  ... Except, wouldn't you need to listen for the corresponding X
> > > > event in a separate thread?  Maybe not ... I'm not really familiar enough
> > > > with X programming to speak with much authority on this particular matter.
> > > No. Why do you think that would be needed?
> > 
> > Because LibGGI doesn't require you to go into some kind of event loop.
> >
> > So it depends a bit on definition. If we just post a regular expose_event,
> > this is more or less unnecessary (at least for the X target), because the
> > target itself could redraw, then.

> ??? ;-)
> You either poll for the expose events, in which case you will
> recieve them, or you don't poll and don't recieve them.
> In neither case a separate thread makes any difference.

Well, I think what he's talking about is having a second thread in the X
target lib that polls for X expose events itself, so they never have to fall
through to the application itself and be processed there before the screen
can be updated.

-=MenTaLguY=-