You could also do this with the Clojure LSP binary

Well, then it is the LSP-organize-imports equivalent which takes into account clj-kondo

unfort the lsp impl doesn't do as good a job as cljr does when it comes to .cljc files

you're right though, that would have been ace

my colleague @U0HJ7CX6H will file an lsp issue about it at some point 🙂

You could maybe run eMacs as a binary

not sure if you're joking haha

i half considered just calling the emacs binary to do the thing 😄 clojure-lsp almost does the right thing, but struggles with namespace reader conditionals, specifically ones that don't use splicing but rather two different require sets for clj and cljs

I wasn't joking :)

😅

hi, refactor-nrepl (clj-refactor's backend) maintainer here 👋 refactor-nrepl is a vanilla clojure library which you can simply invoke however you please. https://github.com/clojure-lsp/lein-clojure-lsp and similar do a nice job as well refactor-nrepl is accurate indeed, but the formatting is a bit wacky for .cljc files. What I do in my daily workflow is run refactor-nrepl first, then a formatter of choice (I happen to use https://github.com/nedap/formatting-stack but anything will do)

(maybe in a future refactor-nrepl will use zprint internally so nothing wacky is spat)

wow didn't know about formatting-stack!

this is amazing

formatting-stack looks amazing! thank you @U45T93RA6

Cheers 🍻 It's quite underpromoted (even though @UHJH8MG6S and me keep it alive), we have some finishing touches pending (like an easier CI setup)

Thanks for sharing formatting-stack. I wanted to try it on our repo but just (require '[formatting-stack.core]) caused a huge CPU usage spike (800% for minutes until I killed the REPL)

@U06BE1L6T probably it's the cost of running clj-kondo analysis in a project for the first time (which is done here https://github.com/nedap/formatting-stack/blob/8b0482d95d22628eeb8991d8503268b650089c17/worker/formatting_stack/background.clj)

I always run f-s in all projects I work on / open (I have dozens of random clones, besides from work) - it doesn't have a generalized perf issue - much the opposite. Perhaps your project is particularly large?

It's not a huge project but it is a bit larger. I simply ran (format!). How does it decide what to analyze?

(taking this to DM to avoid further thread hijacking 🏴‍☠️)