Re: [PATCH 3/5] d3d8/tests: Test detaching adapters.

[Zhiyi Zhang]

On 3/25/20 3:27 AM, Henri Verbeet wrote:
> On Tue, 24 Mar 2020 at 12:22, Zhiyi Zhang  wrote:
> > -
> >    /* Detaching adapter shouldn't reduce the adapter count */
> > -
> >    expected_adapter_count = adapter_count;
> > -
> >    adapter_count = IDirect3D8_GetAdapterCount(d3d);
> > -
> >    ok(adapter_count == expected_adapter_count,
> > -
> >            "Adapter %u: Got unexpected adapter count %u, expected %u.\n", i, adapter_count,
> > -
> >            expected_adapter_count);
> > -
> > -
> >    monitor = IDirect3D8_GetAdapterMonitor(d3d, i);
> > -
> >    ok(!monitor, "Adapter %u: Expect monitor to be NULL.\n", i);
> > -
>
> Not an issue with this patch, but this does raise the possibility that
> some of the earlier checks in this test may not be correct for the
> case where a monitor was detached prior to running the test.

If a monitor is detached prior to d3d instantiation, then it's not enumerated.
Only attached monitors are enumerated during d3d initialization. So this case
could only happen when a monitor is attached before d3d instantiation, and then
the monitor gets detached while d3d is still in use. I think this is also verified by
the earlier checks in this test.

Thanks,
Zhiyi