No numpy.float128 on Windows 64-bit; xferfcn_input_test.py failing with import error #316
Description
xferfcn_input_test.py explicitly imports float128 from numpy, and this fails on Windows 10 64-bit.
From a Windows 10 64-bit machine in a Conda environment:
Python 3.7.3 (default, Apr 24 2019, 15:29:51) [MSC v.1915 64 bit (AMD64)] :: Anaconda, Inc. on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy
>>> numpy.__version__
'1.16.4'
>>> for ftype in [f for f in dir(numpy) if f.startswith('float')]: print(ftype)
...
float
float16
float32
float64
float_
float_power
floating
>>> numpy.longdouble
<class 'numpy.float64'>
>>> from numpy import float128
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
ImportError: cannot import name 'float128' from 'numpy' (C:\Users\rory\Miniconda3\envs\work\lib\site-packages\numpy\__init__.py)
See also failing Slycot test w/ python-control on Appveyor.
This is not very new code; I guess no-one runs the tests on Windows 64-bit?