Re: Bug in kwin (kde2)

["Valeriy E. Ushakov" <uwe ptc spbu ru>]

On Fri, Nov 17, 2000 at 15:25:02 +0100, Matthias Ettrich wrote:

> > I don't know KDE internals, so this might be wrong, still the
> > existing code confuses mask and state and need to be fixed.
> >
> 
> *blush*
> 
> Your are right, of course.
> 
> Thanks for looking that up. Fixed now.

Not really :-)  It seems thoroughly messed...

When the window is already maximized and a request is made to maximize it

    m_client->maximize( Client::MaximizeFull );

is correctly called but this call restores the window (which is
probably a problem with 'maximize' method, but I don't fill like
digging in a huge and completely unfamiliar C++ code base).

And when it comes to setting vertical and horizontal maximization
_separately_ - it's even more fun.

For example I can see this sequence:

set   MAXIMIZE_HORIZ -> max horizontally
set   MAXIMIZE_VERT  -> restores
clear VERT | HORIZ   -> stays restored
set   MAXIMIZE_VERT  -> max full
    // changeState was called with (state = 0xc, mask = 0xc)
    // that is with NET::Max, though only the request is only for vertical 

Looks like some state bits got stale somewhere and fell out of sync
with reality.

PS: And for a bonus: _NET_WM_STATE_SHADED atom is not added to
_NET_WM_STATE when the window is shaded either by the user or by a
programmatic request.

:-(

PPS: I think we're drifting off-topic, so perhaps we can continue in
private e-mail (this message is only BCC'ed to wm-spec).

SY, Uwe
-- 
uwe ptc spbu ru                         |       Zu Grunde kommen
http://www.ptc.spbu.ru/~uwe/            |       Ist zu Grunde gehen