E_NOTFOUND returns the infotech.h value instead of the Win32 error HRESULT

[jeremy-visionaid]

Actual behavior

E_NOTFOUND generates:

internal static readonly winmdroot.Foundation.HRESULT E_NOTFOUND = (winmdroot.Foundation.HRESULT)(-2147479539);

This is the value from infotech.h:

./um/infotech.h:#define E_NOTFOUND _HRESULT_TYPEDEF_(0x8000100DL)

Expected behavior

internal static readonly winmdroot.Foundation.HRESULT E_NOTFOUND = (winmdroot.Foundation.HRESULT)(-2147023728);

Which is the HRESULT for ERROR_NOT_FOUND (um/xamlOM.h and um/devicetopology.h and ./um/mmdeviceapi.h)

i.e.

#define E_NOTFOUND HRESULT_FROM_WIN32(ERROR_NOT_FOUND)

Repro steps

1. NativeMethods.txt content:

E_NOTFOUND

2. NativeMethods.json content (if present):

3. Any of your own code that should be shared?

Context

• CsWin32 version: 0.3.275
• Win32Metadata version (if explicitly set by project):
• Target Framework: net10.0-windows10.0.26100.0
• LangVersion (if explicitly set by project): [e.g. 9]

[jeremy-visionaid]

It's unfortunate that the Windows SDK defines two different values for the same identifier, but I'm not sure if something can be done in Win32Cs or the Win32Metadata to give a warning or error in those cases, and/or disambiguate it for the generator...

[jevansaks]

It seems like a bug, I think most places would expect E_NOTFOUND is 0x80070490.