Re: [PATCH 0/5] MR3607: apisetschema: Add ext-ms-win-dxcore-l1-1-0.

[Rémi Bernon (＠rbernon)]

Rémi Bernon (@rbernon) commented about include/dxcore_interface.idl:
> - uuid(f0db4c7f-fe5a-42a2-bd62-f2a6cf6fc83e),
>
> +]
> +interface IDXCoreAdapter : IUnknown
> +{
> - BOOL    IsValid();
> - BOOL    IsAttributeSupported(GUID attribute);
> - BOOL    IsPropertySupported(DXCoreAdapterProperty property);
> - HRESULT GetProperty(DXCoreAdapterProperty property, SIZE_T *buffer, void *data);
> - HRESULT GetPropertySize(DXCoreAdapterProperty property, SIZE_T *buffer);
> - BOOL    IsQueryStateSupported(DXCoreAdapterState property);
> - HRESULT QueryState(DXCoreAdapterState state, SIZE_T state_size, const void *input, SIZE_T buffer_size, void *output);
> - BOOL    IsSetStateSupported(DXCoreAdapterState property);
> - HRESULT SetState(DXCoreAdapterState state, SIZE_T state_size, const void *in_state, SIZE_T data_size, const void *data);
> - HRESULT GetFactory(REFIID riid, [out, iid_is(riid)] IUnknown **ppv);
> - HRESULT GetAdapterByLuid(LUID adapter, REFIID riid, [out, iid_is(riid)] IUnknown **ppv);
>
> +};

Some of these methods have C++ template wrappers, it would probably be better to have them too.

I understand that it might be difficult, because there's currently no way to include custom interface code, but maybe it's a hint that this should not be written in IDL? (or maybe we need a WIDL feature, idk).

Fwiw I personally don't really mind not adding them, at least from Wine perspective the C++ interface is not very useful, but as you're apparently interested in having this for MinGW too it might be important to have it correct there.