[ruff] Unnecessary rounding (RUF057)

[InSyncWithFoo]

Summary

Resolves #14793.

Test Plan

cargo nextest run and cargo insta test.

[github-actions]

ruff-ecosystem results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

[MichaReiser]

Thanks for following up on this!

Would you mind splitting the PR into two if it isn't too much work:

1. Changes the existing rule
2. Adding the new rule

Two separate PRs have the advantage that we can list the two changes separately in the changelog. It also makes it easier for me to review your PR

[InSyncWithFoo]

@MichaReiser I would prefer not having to do that. RUF057 depends on a RUF046 API (expr_is_strictly_int()), which itself comprises most of the change to RUF046.

[MichaReiser]

Yeah, that's a bit annoying. However, I just skimmed over the changes, and there's too much going on in this PR, which makes it difficult for me and future readers to understand what it is about. That's why I have to insist that you split the PR. I suggest three PRs:

1. Scope change of RUF046
   2.. Improving the integer inference of RUF046
2. Adding the new RUF057

I'd suggest you start with one PR and then submit the next once that is merged. Or you can create stacked PRs, but they can be a bit painful when it comes to rebasing them.

[InSyncWithFoo]

All done. I don't mind a little bit of rebasing, so stacked PRs it is.

[MichaReiser]

Thank you.

Hmm. This is not the rule I expected. From our conversation in #14793 I'd expected a rule that detects unnecessary round calls rather than unnecessary ndigits.

E.g. round(5) should be converted to just 5. The same for round(5, None)

[MichaReiser]

Thanks and happy new year. This is looking great. I've one last comment on where I think the rule should flag a round call but currently isn't. I'm interested in why you chose to not flag that round call site.