From: Nix <nix_at_esperi.org.uk>
Date: Tue, 18 Nov 2003 22:27:41 +0000

On Tue, 11 Nov 2003, BlueChip spake:
> But brings me to a serious question ....without resorting to a compiler for proof:
> what is the result of 0/0 (zero divided by zero)
> I can make it zero, one or 'error' depending on the order of evaluation!

In IEEE754 floating point arithmetic, a NaN results.

(However, the SH doesn't appear to fully support IEEE754 without -mieee;
I'm not sure how its behaviour differs when -mieee is not supported,
but NaNs and denormals are the sort of place where you could expect
to find changes.)

-- 
`Me, I want exploding spaceships and pulverized worlds and clashes of
 billion-year-old empires *and* competently written sentences.'
                                                    --- Matt Austern

Received on 2003-11-18