How I understand Clojure's `apply`

I wrote a big ole’ post on apply, but to save you time I’ll shrink it:

To start:

• Clojure functions work with the AFn class…
• which implements the IFn interface…
• which specificies twenty-two overloads of invoke
• and a method called applyTo

Then:

• applyTo takes a sequence
• and passes it to applyToHelper
• this counts the sequence and
• it passes it into a twenty-two branch switch statement
• which decides how much ISeq.first and ISeq.nexting to do
• to pull out every argument into an invoke

apply uses applyTo with Java interop.

(f '(1 2 3)) will run the single-arity f.invoke('(1 2 3)). (apply f '(1 2 3)) will run the triple-arity f.invoke(1 2 3) which is useful for things like max, min, distinct?.

(max [1 2 3])               ;;=> [1 2 3]
(apply max [1 2 3])         ;;=> 3
(apply max 4 [1 2 3])       ;;=> 4  a.k.a. the "chips"
(min [1 2 3])               ;;=> [1 2 3]
(apply min [1 2 3])         ;;=> 1
(apply min 0 [1 2 3])       ;;=> 0
(distinct? [1 1 3])         ;;=> true
(apply distinct? [1 1 3])   ;;=> false
(apply distinct? 1 [1 2 3]) ;;=> false

That’s it. 🙂