Hacker News new | ask | show | jobs
by Jtsummers 1055 days ago
The problem apparently is that they didn't actually have full type inference and the expressions on the right were given their types (probably a tuple of int x int x cstring) before they attempted the assignment into foo. Now they're using unification (they call it "top-down inference", so I'm guessing it's regular unification like other type inference systems use) so that the expression on the right will have the correct type and be assignable into the variable on the left.
1 comments

I see, I guess you're right.