Re: "full screen" windows

[Bj|rn Englund <d4bjorn dtek chalmers se>]

On Fri, 17 Aug 2001, Philipp Lohmann wrote:

> Am 17.08.01, 13:25:24, schrieb Bj|rn Englund <d4bjorn dtek chalmers se> zum 
> Thema Re: "full screen" windows:
> 
> 
> > In v1.1 of EWMH spec the part about _NET_WINDOW_TYPE says:
> > "This SHOULD be set by the Client before mapping"
> >       ^^^^^^
> > How is this to be interpreted?
> > 1. The property must be set before mapping if it is to be set
> > or
> > 2. The property should be set before mapping but can be set after too.
> 
> I think it is clear: set it before mapping. If you set it afterwards and 
> a WM really reacted on that (didn't try) he would have to redecorate it 
> which will lead to undesirable flickering.

Then one can't change window type without doing unmap/map.
This is like with motif hints which is a problem if you want to 
change decorations on a mapped window.
So having a FULLSCREEN WINDOW_TYPE wouldn't be good.
It should be a WINDOW_STATE.
 

> > _NET_WM_STATE
> >  This property can be changed when the window is mapped.
> >  Is this where there should be a FULLSCREEN_STATE ?
> 
> I think it depends on how you see it: is FULLSCREEN simply maximized with 
> decorations outside the screen or is FULLSCREEN a window type with no 
> decoration and maximum size ? The FULLSCREEN state would be more flexible 
> i think.

I agree that it should be a state.
I see FULLSCREEN_STATE as a window with no decoration and max size and on
top(but other windows should be able to be on top of it).

Fullscreen as a maximized window with deco outside the screen wouldn't be
nice since they would be visible on adjacent workspaces.


/Björn