Should this code be valid?

	var b uint8
	_ = float64(1 << b)

Currently go/types and cmd/compile both reject it.

If the left operand of a non-constant shift expression is an untyped constant, it is first converted to the type it would assume if the shift expression were replaced by its left operand alone

So I believe the error "shift of type float64" is correct because what the spec says we do is

float64(float64(1) << b)

so it's true that you are shifting a float.

Yeah. I guess you're right. It'd be nice to make the error message a bit clearer--it happened to me in the context of a much more complex expression, but I don't really see how to. Closing for now; thanks, @ALTree .

Yes, this is correct: The type of the 1 in (1 << b) is the type it would have if the shift wasn't there. Would be nice to have a clearer explanation in the error but I think that will require some work. Let's leave this for the "ponies" list once we have a compiler that's all neat and clean.

A neat and clean compiler should show up somewhere around Go 1.x. Where I am assuming that this release will be called 1.9, the following one 1.a, then 1.b, and so on. :P