formatter moves type:ignore comment and makes line too long

[DetachHead]

before

def run_pytest(pytester, subprocess):
    result = (
        pytester.runpytest_subprocess 
        if subprocess else pytester.runpytest  # type:ignore[no-any-expr]
    )(*(pytest_args or []))

after

def run_pytest(pytester, subprocess):
    result = (
        pytester.runpytest_subprocess if subprocess else pytester.runpytest  # type:ignore[no-any-expr]
    )(*(pytest_args or []))

E501 Line too long (103 > 100 characters)

playground

https://play.ruff.rs/720b0769-e4d7-430f-8e14-5c7295607d6d

[charliermarsh]

This is an intentional behavior as we exclude pragmas when computing line length in the formatter: https://github.com/astral-sh/ruff/blob/main/crates/ruff_python_formatter/README.md#pragma-comments-are-ignored-when-computing-line-width.

In the next release, we're planning to ignore them when enforcing E501 violations to make the linter consistent with this behavior.

[charliermarsh]

I'm gonna close as fixed by #7692, but won't actually be fixed until we cut the next release.

[jonyscathe]

This is an intentional behavior as we exclude pragmas when computing line length in the formatter: https://github.com/astral-sh/ruff/blob/main/crates/ruff_python_formatter/README.md#pragma-comments-are-ignored-when-computing-line-width.

The example given there says that "Ruff will also avoid moving pragma comments" essentially to avoid suppressing more errors than intended.
However in the example above Ruff not only exceeds the line length, but also applies the type ignore to more code than was originally intended (benign here, but could suppress genuine errors in some cases)