Dict unpacking propagates wrong type #412

a, b = {1: "x", 2.0: 1j}

The way iterable unpacking works, a and b should have types based on the keys of the dict, since that's what you get when you convert to the functionally equivalent:

d = {1: "x", 2.0: 1j}
it = iter(d)
a = it.next()
b = it.next()

But, we get:

a is being given the value type, and b probably the first value's type.

Likely related is what happens when changing this to:

a, *b = {1: "x", 2.0: 1j}

b has no type, but should be a list.

Found in #334.