Wrote a little piece on why I’m excited for @biomejs’s upcoming type inference feature:
https://arendjr.nl/2024/01/why-im-excited-for-biomes-type-inference
#typescript in #rust #dev #tooling
Shoutout to @ematipico for reviewing!
@arendjr Wish (if you have time!)—a feed for your blog: RSS, Atom or JSON Feed. JSON Feed is really simple: https://www.jsonfeed.org/
@rauschma I’ll look into it!
@rauschma there’s an RSS link on the blog page now: https://arendjr.nl/blog/
@arendjr Great, thanks!
@arendjr at the same time, when reading https://www.joshuakgoldberg.com/blog/rust-based-javascript-linters-fast-but-no-typed-linting-right-now/ which you linked to, I felt that it might be better to keep the complexity down by concentrating on your thing? I think it's fine to have one fast linter and then a second slower one, working in the background? Specifically because of the challenge you brought up that, that the goal keeps moving away from you. Maybe their 70% way solves that though. Interesting to follow.
@marlun yeah, using two linters is definitely also an option and some like it that way. But others consider it a no-go, so we’re trying to branch out. Biome’s philosophy has always been that it’s a single tool that covers many needs for web developers, so I think this one fits :)
@arendjr yes, there’s definitely pros in having some parts in a single tool. We had problems recently where prettier and eslint didn’t agree (after an update) on parts of the code base which required configuration to make everything pass. Those kinds of problems shouldn’t happen if it’s all one tool I guess