feat: uncurry 0

[anmonteiro]

fixes #674

[anmonteiro]

@jchavarri @andreypopp this is now ready for review. Note the comment (#674 (comment)) where I highlighted we're not handling @mel.meth with @u0 (those are still inferred to be 0-arity based on unit arguments).

Let me know your thoughts!

[jchavarri]

Why removing the newline here? I found it was useful to have it (see changes in tests).

Edit: I see other messages don't include it. Maybe the newline could be added separately to have some space between the alert and the actual output when calling melc?

[anmonteiro]

This should be ready to merge, but there are a few issues with the PR in the current state:

• we don't have a solution for [@mel.uncurry] -- I only noticed it because there was an uncurried alert emitted for something like t -> (unit -> t [@mel.uncurry]) -> unit = ..`
  • we'd either have to add @mel.uncurry0 or go back to the old behavior where a single unit argument represents 0-arity uncurried application
• we totally glossed over the fact that ReasonML doesn't have a syntax for this... so we won't have a way to write foo(. ()) vs foo(. ) at all
  • this is actually blocking for this PR, so we'll need to put this on the back burner until we figure it out

[jchavarri]

• we totally glossed over the fact that ReasonML doesn't have a syntax for this

just for doc purposes, found this issue: reasonml/reason#2357