[Ogle-devel] Re: KWin and Ogle

[Havoc Pennington]

On Thu, Jan 16, 2003 at 12:08:05PM +0100, Bj|rn Englund wrote: 
> Ogle already does this :)

Perfect, good news.
 
> STATE_FULLSCREEN works with metacity, but I have found some problems
> that might need the old method (unmap/motifhints/map).
> What happens when Xinerama is in use? 
> STATE_FULLSCREEN will cover all screens? (most people want only one screen)
> (this should be configurable, but is it in all wms?)
> The same problem when xf86vidmode is used to change resolution.
> STATE_FULLSCREEN will make the window cover the whole screen, not
> the smaller viewport that one wants. (also configurable?).
> So for these to cases one might need to use motifhints and positioning
> from the program.

If people want to configure this per-program, we should probably add
some WM spec support for it (such as an additional state,
FULLSCREEN_ALL_XINERAMAS, or something like that). Or alternatively,
just say you have to use a "matched windows" type of feature in the
WM to configure it per-window-class.

However if it's just a matter of "doing the right thing" rather than
users wanting per-app different behavior, I would consider it a window
manager quality of implementation issue. I mean, if a WM maximizes
across all Xineramas, that would be really broken, but you would not
try to work around it by changing every app; you'd fix the WM. Same
with fullscreen.

FWIW Metacity is at least intended to do fullscreen to the current
screen and to a single Xinerama monitor, and as far as I know it does
work properly. If not, please let me know. Feel free to close any user
complaints about fullscreen behavior with metacity as "metacity's
fault, go away" and send the users after me instead, assuming you're
using the FULLSCREEN hint. If you aren't using the hint, then I'll 
do "Ogle's fault, go away" from my end. ;-)

If we need spec extensions to get the right behavior, please bring
them up on wm-spec-list, we do actively add stuff to the spec.

Havoc