According to mypy_primer, this change has no effect on the checked open source code. 🤖🎉

Btw, I've found many more missing type aliases in other modules via my custom query.

Are you interested in annotated them?

What benefits does this PR bring?

There are some differing opinions on exactly when TypeAlias should be used in stubs -- e.g. #8256 (comment)

I've just seen @srittau's comment on his vision of TypeAlias. Mine is completely different.

I see TypeAlias as a way to say: these things are both types and they are the same thing.

Consider these lines:

x = 1
def some(): ...
class T: ...

y = x
other = some
T1 = T
T2: TypeAlias = T

Here we can notice that y, other, and T1 are assignments. But, we cannot be sure what their values are (until we go and see, it is even more complicated when different name schemes are used like error = _SomeError).

But, TypeAlias gives us strong guarantee: these two are just different names for a single type.

But, feel free to close if this is not something you want, I don't have a strong opinion.

I'd also be hesitant to make a change here unless there's a clear problem being fixed.

Let's close this for now