(Free42) accuracy of LN1+X

[Albert Chan]

Perhaps it is implemented with binary128 version log1p ?
Example:

>>> from gmpy2 import *
>>> mpfr('1e-99',113)       # binary128 precision = 113 bits
mpfr('9.99999999999999999999999999999999925e-100',113)

Free42:

1e-99 LN1+X     → 9.999999999999999999999999999999999e-100