Re: Pending 1.2 stuff

[Matthias Clasen <maclas gmx de>]

On Sat, 2002-08-10 at 04:30, Havoc Pennington wrote:
> 
> Matthias Clasen <maclas gmx de> writes:   
> > only minor changes to integrate better with the rest of the spec.
> >
> 
> Looks good.
>   

I will commit that part of the patch in a few days, unless somebody
raises objections.


> > - _NET_WM_STATE_FLOATING 
> > 
> > this is a bit differs from Havocs _NET_WM_STATE_STAYS_ON_TOP /
> > _STUCK_TO_GLASS. I think you can get both intended behaviours by
> > combining floating state with the right window type:
> > _NET_WM_STATE_STAYS_ON_TOP=_NET_WM_WINDOW_TYPE_NORMAL+_NET_WM_STATE_FLOATING
> > _NET_WM_STATE_STUCK_TO_GLASS=_NET_WM_WINDOW_TYPE_DOCK+_NET_WM_STATE_FLOATING
> 
> Maybe it's a good idea to explicitly mention this in the spec. However
> it doesn't 100% make sense to me; note that the onscreen keyboard
> needs to be above FULLSCREEN windows and splash screens, while a dock
> would normally be below fullscreen windows and splash screens, for
> example.
> 
We could make a special exception and say that floating docks float
above everything...

> It's possible we should break down and add
> _NET_WM_WINDOW_TYPE_ONSCREEN_KEYBOARD since the onscreen keyboard is
> such a totally bizarre app. Or just leave it as
> _METACITY_ONSCREEN_KEYBOARD since metacity is probably the only WM
> that's likely to get accessibility-ified in the near future anyway.

but I think you're right that onscreen keyboards deserve a separate
semantic type, if only to allow WM to implement the right focus
behaviour for them. I don't see how accessibility of the WM itself
relates to this, however. _NET_WM_WINDOW_TYPE_ONSCREEN_KEYBOARD gets
my vote. 

Do you agree with replacing STATE_STAYS_ON_TOP with FLOATING with the 
semantics given in my patch ? 

If so, do you think we should add the opposite _NET_WM_STATE_DROWNING
(?) which would hint the wm to keep the window below other windows of
the same semantic type ?


Matthias