np.rec.fromarrays(...) may fail if resulting array is going to be empty

[proste]

Creation of np.recarray via np.rec.fromarrays(...) fails if provided arrayList are empty lists and dtype is non-trivial (containing sub-arrays).

In this particular case the function call has enough information to create empty array equivalent to output of e.g., np.empty(shape=0, dtype=dtype).view(np.recarray).

Reproducing code example:

import numpy as np

np.rec.fromarrays([[], []], dtype=[('foo', np.int32, 2), ('bar', np.int32)])

Error message:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-2-e2e53c33181b> in <module>
----> 1 np.rec.fromarrays([[], []], dtype=[('foo', np.int32, 2), ('bar', np.int32)])

<environment_dirname>/lib/python3.8/site-packages/numpy/core/records.py in fromarrays(arrayList, dtype, shape, formats, names, titles, aligned, byteorder)
    636         testshape = obj.shape[:obj.ndim - nn]
    637         if testshape != shape:
--> 638             raise ValueError("array-shape mismatch in array %d" % k)
    639 
    640     _array = recarray(shape, descr)

ValueError: array-shape mismatch in array 1

Numpy/Python version information:

>>> import sys, numpy; print(numpy.__version__, sys.version)
1.18.1 3.8.1 (default, Jan  8 2020, 23:09:20) 
[GCC 9.2.0]