[Spice-devel] [PATCH spice-gtk 1/4] gtk: add spice_main_request_mouse_mode()

[Victor Toso]

On Wed, May 04, 2016 at 01:01:47PM +0200, Marc-André Lureau wrote:
> Hi
> 
> On Mon, Nov 16, 2015 at 4:34 PM, Victor Toso <lists at victortoso.com> wrote:
> > Hm.. I would either prefer that mouse mode did not change if it can't
> > because the function and its documentation suggested that it will only
> > try to change the mouse mode.
> 
> I am testing this series again, and trying to understand your issue. I
> am still not able to reproduce it though. I can only imagine the guest
> is having vmport enabled (check Xorg.log "for VMWare VMMouse"). Spice
> expects server mode to be working in general (whereas client mode
> depends on having a tablet + agent running). So there is no easy check
> to know if the server mode should work, as it should. We could have
> qemu tell if vmport is enabled, but that would be a configuration
> issue anyway since there would be no way to make server mode work
> correctly when vmmouse abs/rel mismatched with what the client/server
> selected. You could also have multiple "pointers" attached to the same
> input source in qemu, one using absolute and the other relative...
> 
> One can conclude that spice is incompatible with vmmouse (since we
> don't detect vmmouse configuration mismatch and can easily get stuck
> mouse, also vmmouse is redundant with the tablet we use), so I will
> send  a patch to qemu to disable it by default when spice is used and
> error when enabled.
> 
> 
> -- 
> Marc-André Lureau

Tested now with a F24 guest machine. It works fine with fedora 23 qemu
and upstream client. vmport is off and tablet input is disabled (it
should make no difference as you have mentioned before).

The server-side mouse does not work if 'grab mouse in server mode' is
not checked, is it expected? If yes, we might combine those options then
to make the pointer always be grabbed on server mode.

Can you rebase and send the patches again?

Cheers,
  toso