Revert #893: Avoid setting a Call as a base for classes from six.with_metaclass()

[jacobtylerwalls]

Steps

• For new features or bug fixes, add a ChangeLog entry describing what your PR does.
• Write a good description on what the PR does.

Description

#893 intended to fix a false positive for unused-import of six in a pattern like class Thing(six.with_metaclass(meta, ancestor)) by making the nodes.Call containing six.with_metaclass into a base. Then, this violated assumptions elsewhere, leading to the crash in pylint-dev/pylint#5935.

#893 was itself a partial reversion of #841, a fix for #713. #713 is essentially the same issue as pylint-dev/pylint#5935. So what I'm suggesting is that we revert to #841, and then find another way to deal with the unused-import in pylint, if at all.

(I'd much rather have a false positive than a crash. And the class of false positives is larger than anything to do with six.with_metaclass anyway: see pylint-dev/pylint#1630. So we shouldn't treat six as more special.)

Let's not backport this, to give us time to make a decision about what to do in pylint with pylint-dev/pylint#1630.

Type of Changes

	Type
✓	🐛 Bug fix

Related Issue

Refs pylint-dev/pylint#5935

[coveralls]

Pull Request Test Coverage Report for Build 2503836953

• 1 of 1 (100.0%) changed or added relevant line in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.0008%) to 92.24%

---

Totals
Change from base Build 2493049607:	0.0008%
Covered Lines:	9367
Relevant Lines:	10155

---

💛 - Coveralls

[hippo91]

@jacobtylerwalls @Pierre-Sassoulas i'm ok with this PR.
It is indeed much preferable to have a false positive than a crash.
Sorry for #893. A bad answer to the initial problem.