It's sweet that IFn is a protocol in ClojureScript

reinvoke.cljs

(extend-type js/RegExp
  cljs.core/IFn
  (-invoke ([this s] (re-matches this s))))

(#"foo.*" "foobar") ;=> "foobar"
(#"zoo.*" "foobar") ;=> nil

(filter #".*foo.*" ["foobar" "goobar" "foobaz"]) ;=> ("foobar" "foobaz")

@fogus Interesting! Could you point me to the source for this so I can see how it's done? I ack'd the ClojureScript source for extend-type and extend but didn't find the magic. :)

@mjijackson the behavior @fogus mentions is incidental, and is a result of the way :keywords are implemented atop JavaScript strings. That said, I would classify this behavior as seekrit and not sweet because it depends on implementation detail.

But then, maybe seekrit things are sweet?

This Old Gist™

Despite the fact that this is seekrit I still point it out to clojure journeyfolk because I think it teaches deep Clojure by way of radical ClojureScript in one beautiful motion.

Nowadays, no longer need to hide seekrits. Keywords and symbols are actually types that support IFn:

https://github.com/clojure/clojurescript/blob/r1.7.228/src/main/cljs/cljs/core.cljs#L2941-L2945
https://github.com/clojure/clojurescript/blob/r1.7.228/src/main/cljs/cljs/core.cljs#L945-L949

@mfikes I can understand keywords, but why are symbols functions of collections? That seems like a juicy one to leave open for users.

Edit
Nevermind, answered my own question. Clojure does the same thing, which must have been the rationale. Fortunately regexes and many other atoms are available for IFn extension 😈