Drop BanLevel in favor of Optional<BanReason> #16038

BanLevel was a very narrow concept which served a limited purpose. Removing it removes a layer of indirection.

Alternative to #16018 Fixes #15929

<!--e57a25ab6845829454e8d69fc972939a-->

The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.

<!--174a7506f384e20aa4161008e828411d-->

Conflicts

Reviewers, this pull request conflicts with the following ones:

• #16018 (banman: BanLevel enum by dakk)

If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first.

This PR gives no rational at to how it improves the software for the users and is incompatible with the considered future use of creating using banning as a priority. The use of an Optional type needlessly complicates and embrittles the code (e.g. having to insert additional != nullopt checks).

Concept NACK. I just posted in #15929 that I think all this discussion around style/refactor of code whose behavior is not being changed is a waste of effort. Let's revisit the implementation when the behavior involved gets extended or improved instead.

Thanks guys. I don’t believe it’s categorically a waste of effort, as if no cleanup is done along the way, even a consistently well-implemented incrementally advanced project will accumulate significant and problematic technical debt over time, or will fail to apply improvements that would otherwise reduce the risk of issues in the future - a minor example of this is moving to switches that error when a new case is added but not handled.

But I’m sympathetic to the argument that this is not the change to prioritize wrt time and attention, given its impact.

Greg, I’ll take your point about refocusing on user benefit.