Clojurians Log v2

Made an (unedited, improvised) video on how we (at work) solve our deps.edn monorepo issues with a babashka script. https://youtu.be/PIwKYJh3r-Y

TIL

(#(+ % (#(- % 2) 3)) 3) => 4

I thought clojure didn't allow nesting anon functions 😱

babashka does ;)

There is an issue for this: https://github.com/borkdude/edamame/issues/76

clj-kondo should probably warn about this too

https://github.com/clj-kondo/clj-kondo/issues/636

I dragged that clj-kondo issue all the way to the top now

I’m writing a small Clojure primer and had just mentioned “function literals can’t be nested”, then evaluated my example for it. (Which to my surprise wasn’t marked by the linter.) Only, I was using a babashka REPL. 😃

Maybe discourage those fn literals a bit, since they are ugly anyway

This is obviously subjective ;)

But I think I could live without them

It needs to be short, don’t have room for judging, 😃

make sense ;)

They are quite unnecessary, though, I agree. I’m mainly mentioning them as part of mentioning the dispatch character, and also so that you will recognize them when reading Clojure code.

I’m moving more towards “named” lamdas myself, actually.

We could have an optional linter for disallowing fn literals, but I'm not sure if I would go so far as to use it myself

Is Clojure’s non-nesting rule a sort of opinion, you think?

I think it's mostly to avoid confusion, I don't see another reason

#(foo % #(foo %))

this expands into

(fn [%1] (foo %1 (fn [%1] (foo %1))))

which is just valid clojure

user=> '#(foo % #(foo %))
(fn* [%1] (foo %1 (fn* [%1] (foo %1))))

I guess nesting fn literals also leads to the problem of shadowed locals

which you can have with normal fns as well, but maybe not as much