Release branch for 2.17.5

[Pierre-Sassoulas]

What's new in Pylint 2.17.5?

Release date: 2023-07-26

False Positives Fixed

• Fix a false positive for unused-variable when there is an import in a
  if TYPE_CHECKING: block and allow-global-unused-variables is set to
  no in the configuration.

  Closes False positive unused-variable for TYPE_CHECKING imports #8696

• Fix false positives generated when supplying arguments as **kwargs to IO
  calls like open().

  Closes False positive unspecified-encoding when passing encoding as kwarg #8719

• Fix a false positive where pylint was ignoring method calls annotated as
  NoReturn during the inconsistent-return-statements check.

  Closes False positive for inconsistent-return-statements with method annotated with NoReturn return type #8747

• Exempt parents with only type annotations from the invalid-enum-extension
  message.

  Closes False positive invalid-enum-extension for Enum parent class with annotations #8830

Other Bug Fixes

• Fixed crash when a call to super() was placed after an operator (e.g.
  not).

  Closes Crash using isinstance + super due to Astroid #8554

• Fix crash for modified-while-iterating checker when deleting
  members of a dict returned from a call.

  Closes Crash: AttributeError: 'Call' object has no attribute 'name' #8598

• Fix crash in invalid-metaclass check when a metaclass had duplicate
  bases.

  Closes Crash "Fatal error while checking" with metaclass #8698

• Avoid consider-using-f-string on modulos with brackets in template.

  Closes Avoid consider-using-f-string on modulos with brackets in template #8720.

• Fix a crash when __all__ exists but cannot be inferred.

  Closes The undefined variable '__all__' results in an inference failed  #8740

• Fix crash when a variable is assigned to a class attribute of identical name.

  Closes Setting the identical name between an attribute and a variable causes a fatal error #8754

• Fixed a crash when calling copy.copy() without arguments.

  Closes no-argument call to copy.copy() crashes #8774

Other Changes

• Fix a crash when a nonlocal is defined at module-level.

  Closes Checking incorrect usage of the keyword 'nonlocal' results in fatal errors. #8735

[mbyrnepr2]

👍

[codecov]

Codecov Report

Merging #8887 (2c513b6) into maintenance/2.17.x (c6658ba) will not change coverage.
Report is 1 commits behind head on maintenance/2.17.x.
The diff coverage is 100.00%.

Additional details and impacted files

@@                 Coverage Diff                 @@
##           maintenance/2.17.x    #8887   +/-   ##
===================================================
  Coverage               95.57%   95.57%           
===================================================
  Files                     178      178           
  Lines                   18844    18844           
===================================================
  Hits                    18010    18010           
  Misses                    834      834           

Files Changed	Coverage Δ
pylint/__pkginfo__.py	100.00% <100.00%> (ø)

[Pierre-Sassoulas]

We won't merge this back in the main branch because 3.0 is too far away from 2.X at this point. This would break the pre-commit autoupdate, so we might do a history only merge.

[jacobtylerwalls]

Quick question: from where did this change arise?

When I --generate-toml-config on 2.17.5 I get:

disable = ["raw-checker-failed", "bad-inline-option", "locally-disabled", "file-ignored", "suppressed-message", "deprecated-pragma", "consider-using-augmented-assign", "attribute-defined-outside-init", "invalid-name", "missing-module-docstring", "missing-class-docstring", "missing-function-docstring", "protected-access", "too-few-public-methods", "line-too-long", "too-many-lines", "trailing-whitespace", "missing-final-newline", "trailing-newlines", "bad-indentation", "unnecessary-semicolon", "multiple-statements", "superfluous-parens", "mixed-line-endings", "unexpected-line-ending-format", "fixme"]

[Pierre-Sassoulas]

Applied tox -e docs (probably from tbump)

[jacobtylerwalls]

Interesting. So there's a bug somewhere here, then?

[Pierre-Sassoulas]

Maybe it's based on the local pylintrc ?