Some methods on ID3D11DeviceContext seem to be returning inappropriate Results

[sagacity]

Summary

ID3D11DeviceContext has the RSGetState and IAGetInputLayout methods (and perhaps others) that return their results wrapped in a windows_core::Result. However, unwrapping this result panics like so:

called `Result::unwrap()` on an `Err` value: Error { code: HRESULT(0x00000000), message: "The operation completed successfully." }

It seems like there should not be an Err here, since the HRESULT indicates everything is fine. Additionally, I'm not sure why there is a Result involved here in the first place, since RSGetState and IAGetInputLayout return void.

Creating a reproduction scenario is quite involved, since it requires a window, a swapchain, a Direct3D device and so on. If it's necessary I can whip up a repro scenario, of course! But maybe you can already spot the issue. Let me know if you'd like more details.

Crate manifest

windows = { /* path to master repo */, default-features = false, features = [
"Win32_Graphics_Direct3D",
"Win32_Graphics_Direct3D11",
"Win32_Graphics_Dxgi_Common",
"Win32_Graphics_Gdi",
"Win32_System_LibraryLoader",
"Win32_System_Memory",
"Win32_System_Threading",
"Win32_UI_HiDpi",
"Win32_UI_WindowsAndMessaging",
]}

Crate code

No response

[kennykerr]

An error code of zero - S_OK - just means that the API returned a null pointer value on the ABI so there was no interface to populate the Ok variant of Result.

[sagacity]

Ah, fair enough. So I should probably interpret it as an empty Option, in that case. Thanks for the info!

[kennykerr]

Yes there's #2856 tracking that although it doesn't apply to most APIs which is why its hard to implement in practice.

[sagacity]

Argh, sorry. I did a search but I apparently messed that up. Thanks for the great support, btw.