Re: TAKE_ACTIVITY patch

[Olivier Chapuis <chapuis nosmap lri fr>]

On Fri, Oct 08, 2004 at 04:52:39PM +0200, Lubos Lunak wrote:
> On Sunday 03 of October 2004 20:41, Elijah Newren wrote:
> > On Sun, 03 Oct 2004 13:50:57 -0400, Havoc Pennington <hp redhat com> wrote:
> > > It seems more important to say that the WM must pass through the button
> > > release than to say the client has to wait for it.
> >
> > I don't follow.  I tried writing a workaround which would emulate this
> > behavior for clients that didn't support this spec within the window
> > manager.  In no case did the WM ever get a ButtonRelease event, and I
> > wasn't able to figure out how to get the WM to obtain one (well, I
> > guess I could have grabbed, but I was worried that might break the
> > client apps).  Others appear to have tried to get a ButtonRelease
> > before as well, reading through Gnome bugzilla...
> 
>  I think the client which gets a ButtonPress gets some kind of a grab and will 
> always get the matching ButtonRelease.
>

Yes, I think something like that happen. When I implemented the
TAKE_ACTIVITY patch + my proposal for any applications (i.e., no
application collaboration) I was forced to use XRecord. With
XRecord plus others small hacks I was able to detected XDND and
text selection.

Regards, Olivier