Closed abelcheung closed 2 years ago
That's a shame. Maybe just start with Mapping[Any, ...]
so you don't have to worry about key type invariance for now, and defer a more precise type for later. It's better to have incomplete stubs than no stubs at all.
See #29. I think we can get away with replacing _AnyStr
by just str
.
Sorry, I failed to deliver stuff promised; all attempts to write generic mapping-like protocol or type specific protocol fall flat on my face. Maybe mapping would support covariant keys in future, or I can come up with something useful in future, who knows.