Does the repetition scan stay reasonable on a large diff?

[alexander8708]

Our biggest PRs run to several thousand lines. Does the repetition check rescan the file per element, and is there a way to see which rule produced a fail?

[Herricka]

The window is a sliding counter: analyze.RepetitionWindow keeps a cursor and only walks forward, so the pass is linear in the diff length instead of quadratic. For the 'why', policy.WithRule(name, err) wraps a failure in a RuleError that prints the rule by name - and it implements Unwrap, so errors.Is still matches the underlying cause.