http://bugs.winehq.org/show_bug.cgi?id=10142
--- Comment #28 from David Taylor davidt-wine@yadt.co.uk 2008-09-30 01:54:55 --- I am using gnome -- with compiz. In any event, neither kwin nor metacity exhibit the problem as they handle virtual desktops in an entirely different way -- they don't move the windows to negative positions.
Regardless, I maintain that this is a wine bug because firstly, no other program exhibits the irrational behavior of unmapping itself when certain events happen, merely because it is off screen; secondly, there is already code in there to prevent the unmap occurring if the move off-screen is a result of a (subset of) X11 events. It therefore doesn't make sense to me to go ahead and unmap it anyway if SetWindowPos is called without moving the window, when it is already off-screen & still visible.
So, before I have to install every window manager on the planet, could someone explain _why_ wine is unmapping itself if it gets moved off the screen -- and why it is not obviously a bug? Again, it is wine taking deliberate action to unmap itself, without intervention from the window manager. All the window manager does is move the window off-screen (at this point wine DOES NOT unmap itself). Wine unmaps itself when SetWindowPos is later called for some other reason (e.g. running 'cmdow' for example) WITHOUT moving the (already off-screen) window, and wine then unmaps the window. I don't see what compiz or fmvw are doing wrong here.