The floating point has been SUBSTANTIALLY rewritten. I've incorporated SoftFloat into the FPU emulation to get extended precision floating point. I've also added most of the FPU functionality that's usually only present in B2 when using x86 assembly. The only compromise left is the transcendental functions are still done in double precision. I seem to have cured most of the FPU problems - the scrollbars in OS8.x work, Arashi and Maelstrom play properly, the calculator gives proper answers... I'd appreciate it if people really checked this as well as they can. This was a BIG rewrite of B2 code (for the FPU). This code should also be portable, so it could help folks trying to port B2 to other platforms as well. Given the rewrite in the FPU and other more minor changes to the B2 code that AREN'T platform specific, I've decided to bump the B2 version from 1.0 to 1.1.