1 commit

Author	SHA1	Message	Date
Andrew Garrett	d9fd747f2d	Revert "Adds explicit round-off checking to operations that are sensitive to floating point vs. integer round-off errors.", et al. ( r46683, r46671 ) ... Reimplement, but better! Use BC Math library where possible, for fixed-point math operations. The tolerance aspect has been included by doing math internally at 16 digits, but discarding the last two digits for the purpose of comparison.	2009-02-12 22:29:32 +00:00