Clojurians Log v2

Also kinda worried about the performance implications. a lot of nested extra async/await all over the place probably can't be good performance wise. there can be a lot of IIFE generated and they all then imply one extra promise+await creation

I mean not everything always needs an extra IIFE in all contexts. the analyzer is smart enough to figure out when it actually is, but let for example doesn't always need one.

(let [x (if some-condition (foo) (bar))]
  (using x))

let ret123;
if (some-condition) {
  ret123 = foo();
} else {
  ret123 = bar();
}
x = ret123;
using(x);

would be perfectly reasonable code in many contexts I think

This crossed my mind as well. But I think it might be better to do this as a second optimization issue. Existing code isn’t affected by the async / await feature so no regressions there. New async/await code might not be as optimal as it could be, but optimizing let to get rid of IIFE etc is a much bigger scope. Yeah, we can track if something async is happening inside a potential IIFE and only make it async if it needs to be, probably not difficult, but I’d rather do that as a second round once we have the basics in place, unless @dnolen thinks it’s essential.

Btw, isn't Closure clever enough to optimize some of these?

I wonder how stack traces would look like with a bunch of async functions, probably as scary as core.async ones

Try it out. It looks much better from my experience

Just similar to native JS stack traces when writing async/await

man, do-expressions are still in stage 1: https://github.com/tc39/proposal-do-expressions