Re: checking for Not a Number?

[Glenn Maynard <glenn@...>]

On Thu, Dec 21, 2006 at 01:12:10AM -0500, Joe Smith wrote:
> Directly compairing floating point numbers for equality is *always* a bad 
> idea.

I'd be interested in an assembly output scenario that could cause this
to fail:

bool b() { return true; }
float f() { return b(); }
assert(f() == 1.0f);

-- 
Glenn Maynard