JRPN 15C

Last updated on Oct 5, 2024

* Implement decimal floating point multiplication and division
to replace platform's IEEE binary floating point math (Issue 76).
* Round away from zero to more closely match HP's rounding behavior.
* Use decimal FP where appropriate, including matrix operations and
statistics operations on the 15C.
* Change internal overflow detection architecture to be more robust.
* 15C: Fix complex ABS function
* 15C: Use 1e-11 as minumum pivot in matrix LU decompositon (Issue 123)