[PATCH] Refactor sameNoopInput (determining tail-call legitimacy)

[Stephen Lin]

> Similarly, I think your method of comparing struct elements one "real"
> type at a time and ignoring any structural differences is OK with all
> calling conventions currently implemented, but is there any guarantee
> that this will always be true? Maybe this should call into the target
> lowering info to check that this is OK, as well? (All current targets
> can just trivially implement this hook to return "true", so it won't
> really make a difference in practice, but it at least provides a
> mechanism for documenting what the assumptions are and opting out of
> this behavior if desired.)
>

Ahh, I withdraw this part, since it seems like this behavior is
guaranteed by ComputeValueVTs, which is not only target-independent
but implemented in the very same file :)

Stephen