> During immediate function execution fvwm tries to get exclusive
> control of the pointer (a "grab"). This failed in the given case
> because some other program held the grab. Fvwm retries grabbing
> for a second before giving up (the function is aborted). It seems
> that for whatever reason vmware grabbed control over the pointer
> for a long time, which it really should not do because it blocks
> all other applications, including the window manager. But this is
> not related to the vanishing window.
I believe that VMWare grabs the X pointer when it converts the X
pointer into a virtual machine mouse (or non-mouse, if the virtual
machine isn't using it). The pointer stays grabbed until you manually
tell VMWare to release it or, under some conditions, until you move the
mouse pointer 'outside' the virtual machine's window. I think VMWare
also grabs the keyboard here for the same reason.
(I believe that the VM needs to be using the mouse and have the VMWare
tools installed, at which point the in-VM driver signals VMWare itself
to let go of the mouse.)
All of this is kind of a hack. VMWare wants 100% of the mouse and
keyboard events to go to the virtual machine without your host
environment getting in the way, even if what you're doing normally
has special meaning to the host.
- cks
