Re: [v7 07/10] ddraw7/tests: Test SetColorKey() on cubemap surfaces.

[Henri Verbeet]

On 10 November 2015 at 14:10, Riccardo Bortolato  wrote:
> From: Stefan Dösinger
>
> Signed-off-by: Riccardo Bortolato

I'd prefer it if Stefan would just send this in himself.
> - D3DDEVICEDESC7 device_desc;

...
> - BOOL cubemap_supported;

...
> - hr = IDirect3DDevice7_GetCaps(device, &device_desc);
> - ok(SUCCEEDED(hr), "Failed to get device caps, hr %#x.\n", hr);
> - cubemap_supported = !!(device_desc.dpcTriCaps.dwTextureCaps & D3DPTEXTURECAPS_CUBEMAP);

...
> - if (!cubemap_supported)
> -
>    goto cleanup;

This seems a bit pointless if you're only going to use
"cubemap_supported" once. I think there should be a skip() here.

[v7 07/10] ddraw7/tests: Test SetColorKey() on cubemap surfaces.

[Riccardo Bortolato]

Hi Henri,

I copied the cubemap_supported check from test_surface_lock(), I'm
fine either way, you tell me.

As for whoever is sending the patch, I added this to my serie per
Stefan request (after integrating the cubemap check). No problem if he
needs to send it though.

Ciao,
Riccardo

2015-11-12 11:06 GMT+01:00 Henri Verbeet :
> On 10 November 2015 at 14:10, Riccardo Bortolato  wrote:
> > From: Stefan Dösinger
> >
> > Signed-off-by: Riccardo Bortolato
>
> I'd prefer it if Stefan would just send this in himself.
> > - D3DDEVICEDESC7 device_desc;
>
> ...
> > - BOOL cubemap_supported;
>
> ...
> > - hr = IDirect3DDevice7_GetCaps(device, &device_desc);
> > - ok(SUCCEEDED(hr), "Failed to get device caps, hr %#x.\n", hr);
> > - cubemap_supported = !!(device_desc.dpcTriCaps.dwTextureCaps & D3DPTEXTURECAPS_CUBEMAP);
>
> ...
> > - if (!cubemap_supported)
> > -
> >    goto cleanup;
>
> This seems a bit pointless if you're only going to use
> "cubemap_supported" once. I think there should be a skip() here.

[v7 07/10] ddraw7/tests: Test SetColorKey() on cubemap surfaces.

[Henri Verbeet]

On 12 November 2015 at 11:57, Riccardo Bortolato  wrote:
> I copied the cubemap_supported check from test_surface_lock(), I'm
> fine either way, you tell me.

I figured. I don't think the construction in test_surface_lock() is
particularly great either, but at least it uses cubemap_supported more
than once.