Discussion about two new X extension proposals

[Jim Gettys]

On Wed, 2005-05-11 at 12:01 -0400, Adam Jackson wrote:
> On Wednesday 11 May 2005 07:41, Buola Buola wrote:
> > > Basically, the problem is that Wine has a lot of different connections to
> > > the X server (from what I remember, one per Win32 thread). Now, if one
> > > wants to restrict the mouse to a window (for games for example), one
> > > really needs to do this from the thread that actually does the input
> > > (otherwise all mouse events will only go to a thread that may completely
> > > ignore them).
> >
> > If this is only needed for a specific project, maybe it would be
> > better to use synthetic mouse events and don't add rarely-used
> > extensions to the server (I have seen some other thread around here
> > looking for such things to be removed), but if more people think it
> > might be useful it looks ok.
> 
> I think there's an important distinction to make here.  Composite, for 
> example, will probably only ever be used by one application per server, and 
> that's okay, that is in fact the whole point.  Likewise if Wine needs an 
> extension to make its emulation better, that's fine too.
> 

To pick a nit: composite is mostly useful by one client at a time *per
window at a given level in the hierarchy*; it is completely plausible,
that once composite is ubiquitous that toolkits may decide to use
composite as well, but at lower levels of the window tree...

But it is certainly the case that arguing compositing managers of the
root windows would be likely bad news ;-).

> The important part is that the extension gets _used_.  It's not about app 
> count per extension, it's about the number of running servers where that 
> extension is active.
> 

If Wine is having particular problems doing what needs to be done in X,
then it behooves us to listen carefully and do something about it...
			Regards,
				- Jim