Re: Window History Placement

[Havoc Pennington <hp redhat com>]

On Thu, Jan 30, 2003 at 01:18:55AM +0100, Matthias Clasen wrote:
> 
> I guess if declare "Support history placement" within the scope of the
> EWMH, I would consider clients who do it themselves broken, just as I
> consider clients broken who "maximize" their windows by resizing to
> screen size. Well, they are less broken if they only use their own
> implementation as a fallback if WM support is not there, so you may have
> a point. But its still broken since I can no longer globally turn off
> history placement by simply switching to a wm which doesn't support it.
> 

We may want to go back to the older thread I started on this - I think
at some point we considered making the process application-driven, 
e.g. the WM would set a property with serialized window state, and the 
app would restore that property, the WM would unserialize the state
and apply it. So that keeps things in the domain of the application
while allowing the WM to define what "current window state"
is. However it has its own problems no doubt, I don't remember the
whole thread.

Anyway, many apps already save/restore their own position, and I
believe that's the de facto reality if we add nothing to the WM
spec. So even right now, you can't disable placement by switching to a
WM that doesn't support it.

Really the question we have to answer is, do we want to go to extra
effort to move this feature into the WM, or do we want to leave it
with the apps as it is now by default.

Havoc