Marc Mueller

The issue seems to have started with #934, the previous commit is fine: https://github.com/PyCQA/astroid/commit/2e8417ffc2285e798ccdef86f743abb75958e2c6. (Tested with pylint `2.8.3`.) @nelfin Would you like to take a look at this?

Cross-posting it here to preserve it for later discussions. -- I'm not convinced we'll see any noticeable improvement with this approach. As far as I'm aware most time is spend...

@nelfin @hippo91 @Pierre-Sassoulas I was thinking again about this issue. At the moment we don't really get anywhere with it. There is #927 which itself is a good change, but...

> @cdce8p: I was/have been fine to wait on sorting the underlying issues before #927, but if you're keen to merge it then I'm happy to put in some extra...

**Update for `not-callable (2)`** ```py data = { 'abc': None, } data['abc'] = lambda: print("Callback called") data['abc']() # false-positive `not-callable` ``` `data['abc']` is inferred as `None`. The reason seems to...

I've already spend some time working on this exact issue 😅 The next step is to deprecate additional default values for arguments in `postinit`. I.e. don't allow a default value...

@DanielNoord You don't need to fix the dataclass brain to fix the issue. It's much easier. See (2) here: https://github.com/PyCQA/pylint/issues/5321#issuecomment-971055235. Just haven't had the time to prepare a PR with...

(2) should be addressed even if you find the issue 😄

> I think we should introduce aliases to match what the `ast` does (without changing the API we have, only adding aliases) That will get confusing quickly. These inconsistencies are...

> Don't we unnecessarily complicate things with this change. Yes and not. The idea was to do `astroid 3.x` after all major changes are done. Just so we keep the...