continuing #839: Extend attrs brain to support provisional APIs

[jstriebel]

Description

This PR continues the original PR #839, which seems to be stale. Only thing missing are formatting and obeying the linter.
From the original PR:

With attrs 20, new "provisional" APIs were added (see https://www.attrs.org/en/stable/api.html?highlight=field#provisional-apis).

This PR adds support to those new functions in the existing attrs brain.

Type of Changes

	Type
✓	✨ New feature

Related Issue

• related to Unsupported object typing when using attr.ib with default #656

[Pierre-Sassoulas]

The coverage is expected to fail don't worry about it coverall is down right now.

[jstriebel]

@Pierre-Sassoulas Thanks for the fast review! I adapted the formatting, CI is green now, but needs an additional OK from a maintainer.

[jstriebel]

@Pierre-Sassoulas Is there anything blocking here? Or could you merge this PR? Thank you!

[Pierre-Sassoulas]

Hey @hippo91 could you review this, please ? I'm not confident enough to merge astroid brains change :)

[hippo91]

@jstriebel thanks for this PR. Everything looks ok. As i understand the new attrs API, define, mutable and frozen are in fact attr.s with specialization of the default arguments.
attrs.field is an alias of attrs.ib. So the modifications proposed here are ok for me even if i wonder about the utility of field and attrib names without attrs prefix.

[jstriebel]

As i understand the new attrs API, define, mutable and frozen are in fact attr.s with specialization of the default arguments.

Exactly, this only adds some new decorator- and function-names that were added with the next-generation API (indeed with different defaults than before), the main mechanisms stay exactly the same.

[jstriebel]

@Pierre-Sassoulas @hippo91 Is there anything missing to merge this? Or do you track this via the milestones and merge it later? Sorry to be bothering 🙏

[Pierre-Sassoulas]

Thank you for the remainder, let's merge that in 2.8.1 as it's finished earlier than I thought :)