I’m working on a small swift-format lint addition and would like design feedback.
The proposed flag is:
swift-format lint --non-fixable-only
It emits only findings that swift-format format cannot fix. The goal is to support workflows that run formatting separately and only want lint output for issues requiring manual attention. Whitespace warnings can get especially noisy while working in an IDE.
The implementation tracks fixability per finding (not per rule) because some rules have both fixable and non-fixable cases.
This use case seems not too esoteric to me, so I'm wondering if others have run into similar use cases and if this is the right interface to expose for addressing this.