Evaluating limits of hyperu function #512
Description
mpmath version: 1.1.0
For certain inputs, e.g. hyperu(1, 2, inf), the implementations for mpmath, scipy, and gsl all agree on the output (0 in this case). However, when the first parameter is a zero, mpmath's implementation diverges from the others (which consistently return 1) by returning a NaN. On the other hand, the version in the fp namespace evaluates to 1, matching the other libraries
Python 3.6.8 (default, Oct 7 2019, 12:59:55)
[GCC 8.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> from mpmath import *
>>> from scipy import special
>>> import numpy as np
>>> mp.hyperu(0, 1, np.inf)
mpf('nan')
>>> special.hyperu(0, 1, np.inf)
1.0
>>> special.hyperu(0, 2, np.inf)
1.0
>>> mp.hyperu(0, 2, np.inf)
mpf('nan')
>>> fp.hyperu(0, 2, np.inf)
1.0
>>> fp.hyperu(0, 1, np.inf)
1.0