Post

Memory doesn't seem to be getting any cheaper, so I'm working on some tools for debugging allocations made by lint rules in oxlint (native rules only). Here's a sneak peek:

9/10/26, 4:48 pm0 reposts5 likes

Arnaud Barré

@arnaud-barre.bsky.social

It's funny how implementing no cycles in linting is always costly but if you implementing this in a script that's realy cheap

Jeroen Engels

@jfmengels.bsky.social

It's cheap for elm-review too. We visit files in the order of imported first, and to do that we compute a graph (& verify there are no cycles). If it has cycles, then we report just that and no other linting errors, it's a necessary prerequisite for the linter. I think other tools should do the same

Arnaud Barré

@arnaud-barre.bsky.social

So the check is required? I would love for this to be enforced more often in JS (there are weird bugs in Vite HMR with cycles), but this would be a no go for too many codebases

Jeroen Engels

@jfmengels.bsky.social

Yes. It's also enforced by the Elm compiler so it's quite natural. I've mostly seen import cycles create super weird effects in JS. How is it being used in a "good" way in JS/TS codebases?