gh-84644: Fix singledispatch annotation parsing

[johnslavik]

A very practical but more general approach than GH-140255 to fixing annotation parsing in functools.singledispatch and functools.singledispatchmethod.

Fixes issues GH-84644, GH-130827, and GH-143886.

It can be broken if one uses a user-defined alternative implementation of staticmethod or something analogous.
Will break incorrect but working registrees. I haven't investigated stripping Annotated typeforms yet.

Consulting a test which fails with this fix at https://github.com/python/cpython/pull/130309/changes#r2663516538 -- I think that the test is wrong.

[johnslavik]

Hm, sorry, I think I was wrong. I've forgotten that a workaround for any potential (still unlikely) breakage is dead simple -- just .register(the_class) instead. That's what the exception message says. Most of my backward compatibility worries were not significant; please forget them. A breakage could only be problematic with third-party libraries, but I believe it to be extremely improbable that their type annotation-based singledispatch registrations would be incorrect. This is rather poor application programming smell.

[johnslavik]

@pR0Ps, I took your test from GH-143888 and committed it with you as an author.

[johnslavik]

Let's see if merging main will help the CI pass.

Offtopic

I'll ping Sviat to check if he's interested in debugging this if the issue persists.
Overall it is I think his action (all-greens).
Otherwise I'll try.

[johnslavik]

Hold on, I'll also think about the prospect of fixing GH-143535 separately.

This could be simplified with that other issue in mind.