Comparing Floating Point Numbers

There's far more to it than I originally thought. Using epsilon = 0.00001f had been working fine for me, but now I see just how wrong that could've been in certain cases.

Definitely keeping this around for reference.

P.S. Do you want find all float==float? Try PVS-Studio rule: V550. An odd precise comparison. http://www.viva64.com/en/d/0141/