Re: [PATCH 1/2] d3drm: Add stubs for IDirect3DRMDevice methods.

[Stefan Dösinger]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am 2015-07-01 um 01:03 schrieb Aaryaman Vasishta:
> +static ULONG WINAPI d3drm_device1_AddRef(IDirect3DRMDevice *iface)
> +{
> - struct d3drm_device *device = impl_from_IDirect3DRMDevice(iface);
> - ULONG refcount = InterlockedIncrement(&device->ref);
> -
> - TRACE("%p increasing refcount to %u.\n", iface, refcount);
> -
> - return refcount;
>
> +}
> -
>
> +static ULONG WINAPI d3drm_device1_Release(IDirect3DRMDevice *iface)
> +{
> - struct d3drm_device *device = impl_from_IDirect3DRMDevice(iface);
> -
> - return IDirect3DRMDevice2_Release(&device->IDirect3DRMDevice2_iface);
>
> +}

Why are you forwarding Release to d3drmdevice2 but not Addref?

I think we should look again at moving the IUnknown methods to Device3,
so we don't thunk to the middle child and stay consistent with the other
methods. Please send an extra patch before this one that migrates the
code to device3. I remember we already looked at this. If this failed
in some spectacular way please remind me :-) .

Is there a logic behind which method forwards to device3 and which just
returns E_NOTIMPL?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
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=jkwG
-----END PGP SIGNATURE-----

[PATCH 1/2] d3drm: Add stubs for IDirect3DRMDevice methods.

[Aaryaman Vasishta]

Is there a logic behind which method forwards to device3 and which just
> returns E_NOTIMPL?

I'm confused about it too, it's been like that since before. I guess we can
do seperate patches for moving around the thunks from version 3 to version
2 (like I did for frame methods).
There was a missing piece of code in this patch which I've fixed in 112636
(Didn't return version 1 device when that IID was passed).