[pydoclint] Fix SyntaxError from fixes with line continuations (D201, D202)

[MeGaGiGaGon]

Summary

This PR fixes #7172 by suppressing the fixes for docstring-missing-returns (DOC201) / docstring-extraneous-returns (DOC202) if there is a surrounding line continuation character \ that would make the fix cause a syntax error.

To do this, the lints are changed from AlwaysFixableViolation to Violation with FixAvailability::Sometimes.

In the case of DOC201, the fix is not given if the non-break line ends in a line continuation character \. Note that lines are iterated in reverse from the docstring to the function definition.

In the case of DOC202, the fix is not given if the docstring ends with a line continuation character \.

Test Plan

Added a test case.

[github-actions]

ruff-ecosystem results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

[ntBre]

Thanks, this looks reasonable to me. Could the same happen with classes? I noticed blank_before_after_class.rs in the same directory, and the implementation looks similar.

[ntBre]

I do wonder if something like iterating over the LogicalLines could be a better solution here:

ruff/crates/ruff_linter/src/rules/pycodestyle/rules/logical_lines/mod.rs

Lines 61 to 62 in 49d7227

	impl<'a> LogicalLines<'a> {
	pub(crate) fn from_tokens(tokens: &Tokens, locator: &'a Locator<'a>) -> Self {

but I'm not totally sure if that's how it works. That might also skip the diagnostic entirely, which could also be reasonable.

[MeGaGiGaGon]

Thanks, this looks reasonable to me. Could the same happen with classes? I noticed blank_before_after_class.rs in the same directory, and the implementation looks similar.

It looks like the same can happen with the before check (D211) since it works the same, but not the after check since it enforces 1 blank line https://play.ruff.rs/a2fac120-588a-4a1d-9287-0663cd0242e3

[ntBre]

Sorry I left this in kind of an ambiguous state. I'll merge this now and we can follow up on classes and the LogicalLines idea separately (if at all, especially in the LogicalLines case). Thanks again!