useless-suppression false positive with ungrouped-imports

[mthuurne]

Steps to reproduce

Run pylint on the following code:

# pylint: enable=useless-suppression
from pylint import run_pylint
import astroid
from pylint import run_pyreverse # pylint: disable=ungrouped-imports

Current behavior

This message is issued:

4:0: I0021: Useless suppression of 'ungrouped-imports' (useless-suppression)

However, if the comment is removed from line 4, this message is issued:

4:0: C0412: Imports from package pylint are not grouped (ungrouped-imports)

Expected behavior

Since there is an actual message being suppressed, I would expect to get no useless-suppression message.

pylint --version output

pylint 2.0.1
astroid 2.0.1
Python 3.4.6 (default, Mar 22 2017, 12:26:13) [GCC]

[PCManticore]

Thanks, I can reproduce this bug. While useless-supression is inoffensive in the sense that it is not consider towards any score or not considered as an error per se, this bug shows that there is some inconsistency somewhere.

[brycepg]

So this particular check as a quit early conditional before adding the message:

if not self.linter.is_message_enabled('ungrouped-imports', import_node.fromlineno):
      continue

Removing the line fixes the issue. Do we want to remove these guards to allow add_message to go through or disable useless-suppression for that check if locally disabled?

[bersbersbers]

This is still an issue with

pylint==2.5.0
astroid==2.4.0

on Python 3.8.2.

I believe my example code is functionally identical to the OP's:

"""False-positive useless-suppression."""
# pylint: enable=useless-suppression
from numpy import array
from pandas import DataFrame

if True and True:
    from numpy import zeros  # pylint:disable=ungrouped-imports

print(array, DataFrame, zeros)

This code gives

bug.py:7:0: I0021: Useless suppression of 'ungrouped-imports' (useless-suppression)

If you remove # pylint:disable=ungrouped-imports, you get

bug.py:7:4: C0412: Imports from package numpy are not grouped (ungrouped-imports)

[bersbersbers]

Please know that the example in #2366 (comment) is functionally different because according to #4212 (comment), the ungrouped-imports is actually unexpected there, so the useless-suppression is expected. I filed #4360.