On Thu, 2015-07-16 at 18:20 -0400, Marc-André Lureau wrote: > Hi > > ----- Original Message ----- > > Virt-viewer creates displays for all potential displays, including ones > > that are currently disabled. So it's incorrect to automatically set the > > display to enabled in this function. For some reason this doesn't cause > > disabled displays to accidentally become enabled, presumably only > > because the displays don't have any dimensions set. But it does cause a > > lot of unnecessary calls to update_display_timer() which sends down new > > monitor configurations to the server. > > --- > > src/spice-widget.c | 1 - > > 1 file changed, 1 deletion(-) > > > > diff --git a/src/spice-widget.c b/src/spice-widget.c > > index 59f9792..5571601 100644 > > --- a/src/spice-widget.c > > +++ b/src/spice-widget.c > > @@ -2434,7 +2434,6 @@ static void channel_new(SpiceSession *s, SpiceChannel > > *channel, gpointer data) > > mark(display, primary.marked); > > } > > spice_channel_connect(channel); > > - spice_main_set_display_enabled(d->main, get_display_id(display), > > TRUE); > > return; > > Have you tested the behaviour with clients relying on resize-guest? Who is enabling the display then? Yes, I have tested with clients that user resize-guest. I have also tested on guests that don't have a vdagent running. The display is being enabled when we receive the very first monitors-config message from the server. Jonathon _______________________________________________ Spice-devel mailing list Spice-devel@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/spice-devel