False-negative E1101 no-member for function parameter

[bersbersbers]

Is your issue fixed on the preview release?: pip install pylint astroid --pre -U --> no

Steps to reproduce

"""False-negative no-member."""


def _func(param=True):
    print(param.name)

Current behavior

No error.

Expected behavior

bug.py:5:11: E1101: Instance of 'bool' has no 'name' member (no-member)

pylint --version output

pylint 2.6.0
astroid 2.4.2
Python 3.8.6 (default, Sep 29 2020, 15:36:55) 
[GCC 7.5.0]

[Pierre-Sassoulas]

Seems like it's working as intended, boolean do not have a name attribute. Am I missing something here ?

>>> b = True
>>> b.name
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: 'bool' object has no attribute 'name'

[bersbersbers]

You may be confusing this with #4047, which is about a false positive. This one is about a false negative. As you rightly say, the code does not run without error, so pylint should complain.

[DanielNoord]

DOesn't this build on either type inference or control flow? What if param is IHaveANameAttribute | bool? This seems like it fits mypy more than it fits us.

[KennyChenBasis]

What about the related

def _func(param : bool):
    print(param.name)

? This also has no no-member emitted (let me know if I should create a new issue for this).

pylint 2.15.5
astroid 2.12.12
Python 3.8.10 (default, Jun 22 2022, 20:18:18) 
[GCC 9.4.0]

[Pierre-Sassoulas]

We could raise a no-member for the function that has True as a default param (original example), but we're probably never going to do anything with type information alone (@KennyChenBasis example). mypy does that and it does it well and faster than us. Even #4813 which is our official "become a discount mypy" issue is about taking the type information into account "when inference fail" not all the time.

[clavedeluna]

What's the agreement here? It seems like the PR that implemented the fix here was closed, so should this issue be closed as won't fix?

[Pierre-Sassoulas]

I think we're going to have too much false positive if we don't wait for #4813, I'll add the blocked label.