np.einsum will not accept NoneType for 'out' parameter

np.einsum raises an error when NoneType (‘None’) is passed to the ‘out’ keyword argument. This behavior is not in line with the expected behavior for the out keyword argument of ufunc methods.

Reproducing code example:

import numpy as np
np.einsum('i,j',[1],[2],out=None)

Error message:

Traceback (most recent call last):

  File "<ipython-input-2-32138e47a353>", line 1, in <module>
    np.einsum('i,j',[1],[2],out=None)

  File "<__array_function__ internals>", line 6, in einsum

  File "einsumfunc.py", line 1356, in einsum
    return c_einsum(*operands, **kwargs)

TypeError: keyword parameter out must be an array for einsum

Numpy/Python version information:

1.18.1 3.7.5 (default, Oct 31 2019, 15:18:51) [MSC v.1916 64 bit (AMD64)]

[WarrenWeckesser]

@eric-wieser, you already fixed this, in https://github.com/numpy/numpy/pull/15256! 😄

Should we add a unit test to ensure it stays fixed?

[sgasse]

Is this still an issue? I can reproduce the bug in an older version like 1.17.2, but not on the latest master:

>>> import numpy as np
>>> np.__version__
'1.19.0.dev0+0b9b176'
>>> np.einsum('i,j',[1],[2],out=None)
array([[2]])