I suggest priority on rewrite of compose #35, as it would help confine current issue

#35 is done now. still not sure how to approach this. I thought it would be possible with a custom curry function (with additional overload function), but it's tricky. Some people use Proxy for this https://medium.com/dailyjs/haskell-like-composition-in-javascript-6142a2a82821 . Of course this could be fixed more easily on the eval side, but it then wouldn't work without it.

Maybe this problem can be solved much easier by flipping it around: https://strudel.tidalcycles.org?fpNMjYivvAhZ

Bringing discussion from #264.. Maybe we should move all the currying inside the patterns.

So add(3) returns a pattern of functions. Then add(3).mul(.5) composes together two patterns of functions, and add(3).mul(.5).fast(2) speeds up that pattern of functions.

This approach would be a challenge for stuff like this though:
s(3).n(2).every(2, add(n(3)).chop(3) )

as chopping implies changing the structure of the outer s(3).n(2) pattern, which I think the chop is too far away from..

But I think s(3).n(2).every(2, add(chop(3).n(3))) would be ok.

This is working now via #390 , as

note("c3").superimpose(
  add.note(3).velocity(.5),
  add.note(7),
)

or

note("c3").superimpose(
  add(note(3)).velocity(.5),
  add(note(7)),
)

#390 was reverted, due to performance issues discussed in #487, so we still don't have composable functions.

One solution could be proxies. The below is proxify, to wrap around functions that take a single pattern as input, so that they can be composed with pattern methods. Applied to id, we get a hitch function that behaves like a lambda:

const handler = {
  get: function (target, prop, receiver) {
    // chain pattern methods
    if (prop in Pattern.prototype) {
      return((...args) => new Proxy((pat) => target(pat)[prop](...args), handler));
    }
    return Reflect.get(target, prop);
  },
};
const proxify = func => new Proxy(func, handler);

const hitch = proxify(id);

$: s("bd sd").apply(hitch.n("<2 4>").chop(32).jux(rev).fast(2))

Attempt at building this into register #1240