Call me dense, but why on Earth would this raise a Floating-Point exception in any compiler? That seems wildly misleading considering there are no FP operation going on, the integer division is supposed to stay in the integer domain - or return an integer NaN, which I'll admit is a rarity. Am I missing something?