On Thu, Sep 07, 2006 at 12:32:02PM +0200, Sander van Dijk wrote:
> On 9/7/06, Anselm R. Garbe <arg_AT_10kloc.org> wrote:
> >Querying the tree here and there is really painful. You'll
> >notice much more load. And it makes no sense, because we have
> >the infomation implicitely if the data structure is slightly
> >extended. (I expect it can be done in 20 additional lines after
> >all with all the benefits I pointed out).
> >
> >Also, focus recovery for tiled clients won't work with the
> >z-layer strategy and focussing always the master is often enough
> >in the way (at least for me).
>
> Hm... But wouldn't that mess up focus order in tiled mode (I mean,
> when doing a couple of non-clockwise mouse-based focus-changes,
> keyboard based focus-order would become quite random).
>
> I think it would be better to use snext (or whatever its name will be)
> for keeping track of the z-order instead (and pop a client whenever
> it's raised by restack()). This would leave focus-ordering intact, and
> still retain z-order (I'll say it again: mixing client-focus-order and
> client-z-order up is bound to lead to problems).
I think you misundertstand something. The stacking is only used
for focus recovery on arrange(), nothing else. It is not
intended to have effect on focusnext/-prev.
Regards,
-- Anselm R. Garbe ><>< www.ebrag.de ><>< GPG key: 0D73F361Received on Thu Sep 07 2006 - 13:51:25 UTC
This archive was generated by hypermail 2.2.0 : Sun Jul 13 2008 - 14:31:06 UTC