On Fri, 29 Jul 2005, David Tweed wrote:
> On Thu, 28 Jul 2005, Tuomo Valkonen wrote:
>
> > Just some minor fixes and improvements here in this early august
> > snapshot.
>
> Has there been some change in the interpretation of _NET_ACTIVE_WINDOW,
> and if so, is it deliberate?
>
> I have a program which uses this to request that a given one of its
> windows become the active window in its frame and then calls a function to
> move the pointer to the centre of the window (the warpPointerTo() call at
> the end). On ion-3ds-20050607 I've got this working ok (through trial &
> error rather than understanding too much about how X does things :-S ),
> whilst on this latest version it seems to find the frame containing the
> desired window ok (as evidenced by the warpPointerTo() call moving the
> mouse pointer to where it'd be expected to be) but it doesn't actually
> make that window the active one in the frame but keeps the one previously
> active (ie, no change).
>
> Since the actual X code which does this is quite short I've included it
> here in case it was only working for the previous Ion by accident.
I was going to say that I'll bet this change was the problem:
Fri May 27 09:38:20 EDT 2005 Tuomo Valkonen <[EMAIL PROTECTED]>
* Ignore _NET_ACTIVE_WINDOW request by default.
- Can be enabled by setting the ignore_net_active_window winprop to
false.
but the version you said you were running before seems like it should
have been tagged after that change went it. I would see if setting the
winprop for the windows you are testing with "fixes" the problem.
If it doesn't we get to keep looking.
-Etan
P.S. _NET_ACTIVE_WINDOW is really a hint and the window manager is more
than free to ignore the request, so it's generally not so nice to assume
it will have worked and proceed, and unless there's a really good reason
I'd leave the mouse alone, apps moving the mouse whenever they want can
get pretty annoying.
