If you're on Clojure 1.10 or later, that could be "-e" "((requiring-resolve '" and if you're on a recent CLI version you should be able to omit those , but otherwise, yeah, that seems right.

I forgot about requiring-resolve, but I'd like for the repl to start in , which is why I wrote it like that.

Cool, thank you

If I find myself writing that much code in an alias, I tend to put it in a function in .clj file and just do a requiring-resolve invocation on it.

I've never understood why anyone cares what ns the REPL starts in -- it is going to track whatever REPL you're eval'ing code from your editor with anyway?

(and I never type into a REPL -- I don't even have the REPL visible in my editor)

At this point, I've also moved any app start stuff into REPL command snippets in VS Code/Calva, so that I can start a "bare" REPL without it starting any processes -- and then I can control what is started/stopped.

(but if you want all your devs to be able to start the app the same way, it makes sense in a .clj file in the repo so everyone can invoke it that way)

But we have ~20 apps so we tend to start/stop different apps depending on what part of the repo we're working in 🙂

Sadly, we have a somewhat complicated docker set up and need a consistent entrypoint for our devs of varying clojure experience. Good point about the repl, you’ve made me realize that I can remove the in-ns stuff because we run the repl headless (in docker). Should have noticed that already.

Thanks for the input