Yes, you're relying on an implementation detail. See the source:

(defn starts-with?
  "True if s starts with substr."
  {:added "1.8"}
  [^CharSequence s ^String substr]
  (.startsWith (.toString s) substr))

(instance? CharSequence 'foo)
false

I guess if the type hint would have been Object then the function would still have worked, but I'm not sure if the intent is that you're throwing random objects in there. Perhaps better to ask @U064X3EF3

Does that I will get better performance too with an explicit call to str, because I can avoid reflection?

(Just listened to Alex’s London Clojurians talk about Clojure performance, was really interesting)

Well, there isn't any reflection going on, I guess .toString is just resolved to Object's toString. It's a strange edge case

I tried testing quickly, added (set! *warn-on-reflection* true) to the top of my ns, didn’t get any warnings.

Thanks for the help! 🙏

The intent here is that you pass it a CharSequence like a String

The toString is necessary because not all CharSequences are Strings

Not sure if that answers your question

yep, that's how I understand it too. so throwing in a symbol isn't the intent, hence the clj-kondo warning is legit

> Not sure if that answers your question It does! 🙏