as_strided fails on dtypes with overlapping fields

[eric-wieser]

Caused by dtype.descr being broken by design, and __array_interface__ being broken by association.

cc @ahaldane

>>> dt = np.dtype(dict(
            formats=['<i4', '<i4'],
            names=['a', 'b'],
            offsets=[0, 2],
            itemsize=6
        ))
>>> a = np.zeros(10, dt)
>>> np.lib.index_tricks.as_strided(a)
TypeError: data type not understood

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<pyshell#132>", line 1, in <module>
    np.lib.index_tricks.as_strided(a)
  File "C:\Users\wiese\AppData\Roaming\Python\Python35\site-packages\numpy\lib\stride_tricks.py", line 102, in as_strided
    array = np.asarray(DummyArray(interface, base=x))
  File "C:\Users\wiese\AppData\Roaming\Python\Python35\site-packages\numpy\core\numeric.py", line 501, in asarray
    return array(a, dtype, copy=False, order=order)
SystemError: <built-in function array> returned a result with an error set