Re: [PATCH 3/7] d3d9/tests: Test messages on focus loss (v3).

[Henri Verbeet]

On 6 October 2014 13:42, Stefan Dösinger  wrote:
> - device = create_device(d3d9, device_window, focus_window, FALSE);
> - if (!device)
> - present_parameters.Windowed = FALSE;
> - present_parameters.hDeviceWindow = device_window;
> - present_parameters.SwapEffect = D3DSWAPEFFECT_DISCARD;
> - present_parameters.BackBufferWidth = 640;
> - present_parameters.BackBufferHeight = 480;
> - present_parameters.BackBufferFormat = D3DFMT_A8R8G8B8;
> - if (FAILED(IDirect3D9_CreateDevice(d3d9, D3DADAPTER_DEFAULT, D3DDEVTYPE_HAL, focus_window,
> -
>        D3DCREATE_SOFTWARE_VERTEXPROCESSING, &present_parameters, &device)))

We're just hoping the display resolution isn't already 640x480 here,
right? I'd also rather see this using create_device(). Or perhaps some
variant like create_device_pp() taking a D3DPRESENT_PARAMETERS
structure, although it seems to me that the issue you're running into
is just that screen_width and screen_height aren't very good choices
for the backbuffer dimensions in the general case.

[PATCH 3/7] d3d9/tests: Test messages on focus loss (v3).

[Stefan Dösinger]

Am 06.10.2014 um 15:49 schrieb Henri Verbeet :
> On 6 October 2014 13:42, Stefan Dösinger  wrote:
> > - device = create_device(d3d9, device_window, focus_window, FALSE);
> > - if (!device)
> > - present_parameters.Windowed = FALSE;
> > - present_parameters.hDeviceWindow = device_window;
> > - present_parameters.SwapEffect = D3DSWAPEFFECT_DISCARD;
> > - present_parameters.BackBufferWidth = 640;
> > - present_parameters.BackBufferHeight = 480;
> > - present_parameters.BackBufferFormat = D3DFMT_A8R8G8B8;
> > - if (FAILED(IDirect3D9_CreateDevice(d3d9, D3DADAPTER_DEFAULT, D3DDEVTYPE_HAL, focus_window,
> > -
> >        D3DCREATE_SOFTWARE_VERTEXPROCESSING, &present_parameters, &device)))
>
> We're just hoping the display resolution isn't already 640x480 here,
> right?

Yes. It seems like a reasonable assumption to me nowadays. I preferred not to complicate the tests with checks for the current resolution.
> I'd also rather see this using create_device(). Or perhaps some
> variant like create_device_pp() taking a D3DPRESENT_PARAMETERS
> structure,

I can do that, but you preferred a direct call to CreateDevice over a parameter when the parameter in question was software vs hardware rendering.
> although it seems to me that the issue you're running into
> is just that screen_width and screen_height aren't very good choices
> for the backbuffer dimensions in the general case.

No, I think it is reasonable for the other tests that use fullscreen mode. On some monitors changing to a different mode takes a lot more time than just switching to fullscreen.