Re: [PATCH vkd3d 4/5] tests: Print adapter info on Windows.

[Józef Kucia]

On Thu, Nov 29, 2018 at 4:31 PM Henri Verbeet  wrote:
> > +#ifdef _WIN32
> > +static void print_adapter_info(void)
> > +{
> > - IDXGIAdapter *dxgi_adapter;
> > - DXGI_ADAPTER_DESC desc;
> > - IUnknown *adapter;
> > - HRESULT hr;
> > -
> > - if (!(adapter = create_adapter()))
> > -
> >    return;
> > -
> > - hr = IUnknown_QueryInterface(adapter, &IID_IDXGIAdapter, (void **)&dxgi_adapter);
> > - ok(hr == S_OK, "Failed to query IDXGIAdapter, hr %#x.\n", hr);
> > - IUnknown_Release(adapter);
> > -
> > - hr = IDXGIAdapter_GetDesc(dxgi_adapter, &desc);
> > - ok(hr == S_OK, "Failed to get adapter desc, hr %#x.\n", hr);
> > -
> > - trace("Adapter: %04x:%04x.\n", desc.VendorId, desc.DeviceId);
> > -
> > - IDXGIAdapter_Release(dxgi_adapter);
> >
> > +}
> > +#else
> > +static void print_adapter_info(void) {}
> > +#endif
>
> I don't necessarily care that strongly, but is the #ifdef really
> needed? It seems print_adapter_info() already handles create_adapter()
> returning NULL.

We currently include DXGI headers only when _WIN32 is defined. We
could include DXGI headers unconditionally, but those are not very
useful when we do not have DXGI.