a naive approach for using kernal floating point numbers on C64, while using IEEE float values at application level

right now CC65 does not support floats by itself, this library is a wrapper to the C64 kernal floating point routines and can be used to call these from C. It may serve as a starting point for a proper library once CC65 has been updated to support floating point.

TODO:

• actually use IEEE floats, the current code uses the CBM binary format (truncated to 32bit)
• conversion from/to (un)signed long is missing
• fix atan2
• fix the polynom calculation functions
• implement more missing functions
• improve the tests
• disable the BASIC errors somehow(?)