Re: [PATCH 3/4] d3d9/tests: Verify window style after exiting fullscreen mode. (try 2)

[Henri Verbeet]

On 1 April 2013 02:40, Sam Edwards  wrote:
> I opted not to write ddraw tests. The existing ddraw window style tests
> suggest that DirectDraw doesn't even set WS_VISIBLE when initializing a
> (non-visible) window, which is very odd, so any new tests would have been
> todo_wine anyway. Nonetheless, my changes to wined3d are minimal and
> shouldn't affect ddraw's behavior.

I think it's ok for the tests to be todo_wine, but we do want ddraw
coverage for this in the tests.
> - ZeroMemory( &d3dpp, sizeof(d3dpp) );
> - d3dpp.SwapEffect       = D3DSWAPEFFECT_DISCARD;
> - d3dpp.Windowed         = TRUE;
> - d3dpp.BackBufferWidth  = screen_width / 2;
> - d3dpp.BackBufferHeight = screen_height / 2;
> - hr = IDirect3DDevice9_Reset(device, &d3dpp);
> - ok(hr == D3D_OK, "IDirect3DDevice9_Reset failed with %08x\n", hr);

I think you'll want to use reset_device() here. (Although I also just
notice 766e732fb changed that to use screen_width x screen_height
instead of 640x480, which is perhaps a bit of an unfortunate choice.
It's fine to change reset_device() to use screen_width / 2 x
screen_height / 2, if needed.)

[PATCH 3/4] d3d9/tests: Verify window style after exiting fullscreen mode. (try 2)

[Sam Edwards]

On 04/01/2013 08:16 AM, Henri Verbeet wrote:
> I think it's ok for the tests to be todo_wine, but we do want ddraw
> coverage for this in the tests.

Fine by me. But that will take a few days since I'll have to familiarize 
myself with ddraw first. :)

In the meantime, should I resend this patchset (with the reset_device() 
change included) so that it can be reviewed and committed while I'm busy 
with ddraw (and send the ddraw tests in a separate patch), or is this 
patchset unacceptable until I include the ddraw tests?

Thanks,
Sam

[PATCH 3/4] d3d9/tests: Verify window style after exiting fullscreen mode. (try 2)

[Henri Verbeet]

On 1 April 2013 16:56, Sam Edwards  wrote:
> In the meantime, should I resend this patchset (with the reset_device()
> change included) so that it can be reviewed and committed while I'm busy
> with ddraw (and send the ddraw tests in a separate patch), or is this
> patchset unacceptable until I include the ddraw tests?

I wouldn't go so far as to say it's unacceptable without the ddraw
tests, but I do think it would be preferred to submit them in the same
set. For what it's worth, I don't think the ddraw tests have to be
particularly complicated, you'd essentially just add a
SetCooperativeLevel(..., DDSCL_NORMAL); call and then test that the
style flags are unchanged from their initial values.