Raw strings vs regular strings

[sobolevn]

One can use both r'abc' and 'abc' to represent a simple string.
But, r strings are only required when dealing with \ in your strings.
So, the rule is: if there's no \ in your string, then do not use r

[sobolevn]

This is an inverse rule to WPS342

[fwald]

Hi! I would like to take on this issue if I may?

[sobolevn]

Sure, @fwald! Thanks a lot! Feel free to ask any questions.

[fwald]

When implementing this, the linter complains about this expression in the file
constants.py:

# Used as a special name patterns for unused variables, like _, __:
UNUSED_VARIABLE_REGEX: Final = re.compile(r'^_+$')

Can this be changed to a normal string instead?

[sobolevn]

@fwald yes please! That's what I love about static analysis the most: finding new issues in code that you have seen so many times!

[Mema5]

Hi, I'm working with @fwald on the issue. Enforcing this new rule also breaks tests in the test suite, so I'm working on modifying it.

But I struggle with this error when running pytest tests/test_visitors/test_tokenize/test_primitives/test_string_tokens/test_string_multiline.py:

___________________ test_incorrect_multiline_strings[other_var + {0}-regular_wrapper-r'''\n'''] ____________________

parse_tokens = <function parse_tokens.<locals>.factory at 0x7ffacd7592f0>
assert_errors = <function assert_errors.<locals>.factory at 0x7ffacd759158>
default_options = options(min_name_length=2, max_name_length=45, i_control_code=True, i_dont_control_code=True, max_noqa_comments=10, ne...el=4, max_attributes=6, max_cognitive_score=12, max_cognitive_average=8, max_call_level=3, max_annotation_complexity=3)
primitives_usages = 'other_var + {0}', primitive = "r'''\n'''"
mode = <function regular_wrapper.<locals>.factory at 0x7ffacd5c5598>

    @pytest.mark.parametrize('primitive', [
        '"""abc"""',
        "'''abc'''",
        '""""""',
        "r'''\n'''",
        'b"""some"""',
    ])
    def test_incorrect_multiline_strings(
        parse_tokens,
        assert_errors,
        default_options,
        primitives_usages,
        primitive,
        mode,
    ):
        """Ensures that incorrect multiline strings are forbiden."""
        file_tokens = parse_tokens(mode(primitives_usages.format(primitive)))
    
        visitor = WrongStringTokenVisitor(default_options, file_tokens=file_tokens)
        visitor.run()
    
>       assert_errors(visitor, [WrongMultilineStringViolation])

tests/test_visitors/test_tokenize/test_primitives/test_string_tokens/test_string_multiline.py:79: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

visitor = <wemake_python_styleguide.visitors.tokenize.primitives.WrongStringTokenVisitor object at 0x7ffacd4dbc50>
errors = [<class 'wemake_python_styleguide.violations.consistency.WrongMultilineStringViolation'>]
ignored_types = None

    def factory(
        visitor: BaseVisitor,
        errors: Sequence[str],
        ignored_types=None,
    ):
        if ignored_types:
            real_errors = [
                error
                for error in visitor.violations
                if not isinstance(error, ignored_types)
            ]
        else:
            real_errors = visitor.violations
    
>       assert len(errors) == len(real_errors)
E       AssertionError: assert 1 == 0
E        +  where 1 = len([<class 'wemake_python_styleguide.violations.consistency.WrongMultilineStringViolation'>])
E        +  and   0 = len([])

tests/test_visitors/conftest.py:31: AssertionError

I added a \n in the primitive to be consistent with the new rule, but something is still wrong and I don't understand why. Do you have a clue @sobolevn ?
Do not hesitate if you want more information on the bug.

[sobolevn]

You can use ignored_types= parameter on assert_errors. This happens when old tests are executed on updated visitor.

Or you can refactor old tests. Both cases are fine.