Neperian logarithm computation for compilers that don't implement it
The microtec C compiler for 68040 doesn't implement the natural logarithm fonction.
This project is a C function that operates on a float type (single precision) and returns the natural logarithm with the float type.
The computation is based on the development in series of the ArcTanH function. Reference : wikipedia
Furthermore it computes separatly the log of exponent and mantissa to bound the error in +/- 10^-9, i.e. at a level lower than float quantification error. The justification of the algorithm is provided in the wiki.