Promote most dunder definitions on FunctionModel to ObjectModel

[jacobtylerwalls]

Description

Create DunderCompletionMixin in order to let ModuleModel share the same dunder definitions that were already present on FunctionModel.

Type of Changes

	Type
✓	🔨 Refactor

Related Issue

Refs pylint-dev/pylint#6094

[coveralls]

Pull Request Test Coverage Report for Build 2447909857

• 27 of 27 (100.0%) changed or added relevant lines in 1 file are covered.
• 229 unchanged lines in 11 files lost coverage.
• Overall coverage increased (+0.2%) to 92.13%

---

Files with Coverage Reduction	New Missed Lines	%
astroid/nodes/scoped_nodes/scoped_nodes.py	1	92.48%
astroid/protocols.py	1	91.84%
astroid/interpreter/_import/spec.py	3	98.24%
astroid/rebuilder.py	5	96.9%
astroid/brain/brain_namedtuple_enum.py	6	93.55%
astroid/nodes/node_ng.py	21	93.17%
astroid/raw_building.py	23	91.79%
astroid/inference.py	32	94.05%
astroid/modutils.py	38	82.95%
astroid/bases.py	40	86.59%

Totals
Change from base Build 2367082861:	0.2%
Covered Lines:	9319
Relevant Lines:	10115

---

💛 - Coveralls

[DanielNoord]

Removing this from 2.12 as I'd like to merge #1606 first, which I'm not sure will be put into 2.12.

[DanielNoord]

Okay, I applied your diff. Merging main creates some interesting failures which makes me wonder if we missed something in #1606

I'll need to investigate that.

[DanielNoord]

Hm, it seems that moving these things to ObjectModel has the side effect that we now start inferring some calls to those new attributes. Since attr___init__ is defined as attr___ne__ all calls to __init__ become uninferable.

[jacobtylerwalls]

Since attr___init__ is defined as attr___ne__ all calls to __init__ become uninferable.

🤕

[DanielNoord]

This is getting closer to completion. The only issue is now that whenever we overwrite __eq__ this doesn't resolve correctly anymore.
I'm not sure if this is for all overwrites of __eq__ or only in brains, as only the numpy brain is failing.

I tried to dig a little deeper but I would like to merge the other PRs first before tackling this one again.

[jacobtylerwalls]

Closing as I don't have immediate plans to work on this.

[DanielNoord]

@jacobtylerwalls Could you restore the branch? I actually wanted to complete this 😄

[DanielNoord]

In the spirit of fixing stale PR's:
@jacobtylerwalls Do you have any idea about a possible fix here? The issue I'm running into above has bugged me for weeks/months and I haven't found a good solution for fixing it. I think the PR makes sense but it feels like the way astroid does inference is just flawed with respect to these dunders. Could you have a look and see if you can help out?

[jacobtylerwalls]

Thanks for the suggestion. I think with the time on hand, my priorities for the moment are triage, performance, and #2048. So I'm unlikely to return to this.