Hi,
I share your sentiment, but in this case it is impossible to know, unless we can find it documented somewhere. From a binary signature standpoint, a function with 0 args is identical in STDCALL and CDECL, because it's not mangled in the DLL either (unlike lib), and the only source we have is that header that uses it.
So it doesn't really matter, just that the mismatch looked wrong (either both STDCALL or both CDECL is fine in my eyes, but not an authoritative opinion).
Oh, I see, perhaps I shouldn't have changed it. I'm not sure if I did the right thing here, especially since it's undocumented.
Giovanni had already signed off on the WINAPI version of this but I saw that it didn't get committed so I thought maybe Alexandre saw something wrong that neither one of us saw. Maybe I was being too impatient, but I wanted to also submit the stubs and those depend on this patch.
I made a decision I thought was right based on my understanding of cdecl and stdcall, but I'm still inexperienced in Wine development. I apologize if I came off as disrespectful towards Giovanni.