Recursion error with hyp2f1 on s390x #1046
Description
The continuous integration tests for 1.4.0 are failing in Debian on the s390x architecture (which is big-endian) with the following recursion error:
>>> import mpmath
>>> mpmath.hyp2f1(-5,5,0.5,0.5)
Traceback (most recent call last):
File "<python-input-1>", line 1, in <module>
mpmath.hyp2f1(-5,5,0.5,0.5)
~~~~~~~~~~~~~^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/mpmath/functions/hypergeometric.py", line 254, in hyp2f1
return ctx.hyper([a,b],[c],z,**kwargs)
~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/mpmath/functions/hypergeometric.py", line 229, in hyper
if q == 1: return ctx._hyp2f1(a_s, b_s, z, **kwargs)
~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/mpmath/functions/hypergeometric.py", line 473, in _hyp2f1
v = ctx.hypercomb(h, [a,b], **kwargs)
File "/usr/lib/python3/dist-packages/mpmath/functions/hypergeometric.py", line 126, in hypercomb
v = ctx.fprod([ctx.hyper(a_s, b_s, z, **kwargs)] + \
~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/mpmath/functions/hypergeometric.py", line 229, in hyper
if q == 1: return ctx._hyp2f1(a_s, b_s, z, **kwargs)
~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/mpmath/functions/hypergeometric.py", line 473, in _hyp2f1
v = ctx.hypercomb(h, [a,b], **kwargs)
File "/usr/lib/python3/dist-packages/mpmath/functions/hypergeometric.py", line 126, in hypercomb
v = ctx.fprod([ctx.hyper(a_s, b_s, z, **kwargs)] + \
~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^
... [snip] ...
File "/usr/lib/python3/dist-packages/mpmath/functions/hypergeometric.py", line 229, in hyper
if q == 1: return ctx._hyp2f1(a_s, b_s, z, **kwargs)
~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/lib/python3/dist-packages/mpmath/functions/hypergeometric.py", line 406, in _hyp2f1
if z == 1:
^^^^^^
File "/usr/lib/python3/dist-packages/mpmath/ctx_mp_python.py", line 202, in __eq__
return self.__eq__(other)
~~~~~~~~~~~^^^^^^^
File "/usr/lib/python3/dist-packages/mpmath/ctx_mp_python.py", line 193, in __eq__
return mpf_eq(sval, oval)
File "/usr/lib/python3/dist-packages/mpmath/libmp/libmpf.py", line 455, in mpf_eq
return s == t
^^^^^^
RecursionError: maximum recursion depth exceeded in comparison