Micha Reiser

Hehe, okay, this one is kind of funny, and it's less clear what the intended behavior is. Looking at Black's tests, they are fairly explicit that there should always only...

Hi @taranlu-houzz Sorry, I haven't had time to look into this more.

You can change the `formatArgs` and `lintArgs` to include a `--config indent-width=2`. I'm not sure if the new LSP will have a setting for this that doesn't require a configuration...

I really like the change. Having both key and value in a single item makes the API way more convenient to use. Thanks for doing this refactor.

> I've taken the changes to the pattern-matching nodes back out again, as @dhruvmanila pointed out to me that they might conflict with some changes around that node being planned...

Yeah that makes sense to me. I'm not entirely sure how to implement this yet. I see two approaches: 1. Add special handling for `fmt: skip` and `fmt: off` (and...

Thanks for opening this. I wonder if we could do better once we have a multifile analysis and maintain such a non-configurable allow list. But maybe that's too opinionated?

Would you mind running some hyperfine benchmarks on this change. The `Contributing.md` explains how. I'm asking because hashing `Diagnostic` can be somewhat expensive. I wonder if we should do some...

I reverted the change to the diagnostic message because it floods the ecosystem changes, making it difficult to see the results related to unconditionally emitting the diagnostic.

The changes look reasonable to me but I would like a second opinion, especially if we should gate this behind preview first. @AlexWaygood wdyt?