Re: [PATCH v2 0/4] MR5585: mmdevapi: Implement AudioClockAdjustment_SetSampleRate.

[Huw Davies (＠huw)]

Huw Davies (@huw) commented about dlls/mmdevapi/client.c:
>  clock2_GetDevicePosition
>
> };
>
> +static HRESULT WINAPI AudioClockAdjustment_QueryInterface(IAudioClockAdjustment *iface,
> -
>    REFIID riid, void **ppv)
>
> +{
> - ACImpl *This = impl_from_IAudioClockAdjustment(iface);
> - return IAudioClock_QueryInterface(&This->IAudioClock_iface, riid, ppv);
>
> +}
> -

This looks a bit odd, given that `IAudioClockAdjustment` can be QI'ed from `IAudioClient`.  We need more tests: e.g. can you QI `IAudioClient` from `IAudioClockAdjustment`?  Does QI'ing `IUnknown` from `IAudioClockAdjustment` give the same result as QI'ing `IUnknown` from `IAudioClient` (as it should if it's the same object).  I'm wondering whether our current `IAudioClock` implementation is correct in this regard too.