Parameter typing improvements

[AndreasBackx]

This removes all of the type: ignore mentions in Argument, its subclasses, and the exceptions module. To do so, the following is done:

1. Parameter is now an abstract base class. This just clarifies it cannot be instantiated which was always the point.
2. Parameter.name can no longer be None. It only ever is None when expose_value = False. Instead of having to deal with None and is None everywhere, let's just use "" instead of None when expose_value is False
3. ctx in Parameter.get_error_hint can be None. It always has been like this, but wasn't explicit in typing. Better to be explicit than implicit.

• Add tests that demonstrate the correct behavior of the change. Tests
  should fail without the change.
  • Parameter.name is an implementation detail, it's not in the documentation.
• Add or update relevant docs, in the docs folder and in code.
• Add an entry in CHANGES.rst summarizing the change and linking to the issue.
• Add .. versionchanged:: entries in any relevant code docs.

[Rowlando13]

@AndreasBackx Would you mind rebasing this so it can go out with 8.3.0?

[kdeldycke]

@AndreasBackx I just reviewed this PR following our discussion at #3329 and just realized it is not that big of a change. So I took the liberty to rebase it on top of stable and fix all remaining issues. The only careful part was adapting it to the UNSET/sentinel but in the end it worked well.

I will program this PR for inclusion for the next 8.4.0 release cycle. Can you double-check everything looks OK to you?

[AndreasBackx]

Wow ya, thank you so much. I reviewed, or still am, reviewing the other ones. It's been a while but it looks as expected I think. This PRO looks as how I'd expect it. I'm kind of fuzzy about changing one None -> "" here in a few places but I think that's a fine change for the edge case it might just be and them being internal implementations.

[kdeldycke]

Wow ya, thank you so much. I reviewed, or still am, reviewing the other ones. It's been a while but it looks as expected I think. This PRO looks as how I'd expect it. I'm kind of fuzzy about changing one None -> "" here in a few places but I think that's a fine change for the edge case it might just be and them being internal implementations.

Just refined some comparisons and removed an assert to clean up code.

So let's consider this PR ready to be merged upstream unless a review is raised.

[Rowlando13]

I will fix changes then this is ready for 8.4.0.

[kdeldycke]

I will fix changes then this is ready for 8.4.0.

Just rebased on top of main with appropriate location of the changelog entry. Also solved a merge conflict due to #3371 merge.

This is ready to be merged into main.

[kdeldycke]

Thanks @AndreasBackx for the merge and for removing the left-over branch!