Eric Wieser

@micha2718l, do you think you'll have time to come back to this?

Thanks for coming back to this! I think this comment still applies: https://github.com/numpy/numpy/pull/14843#discussion_r402719113

What are your thoughts on this comment: https://github.com/numpy/numpy/pull/14843#issuecomment-550566562?

Loads just fine on 1.15.4 + 64-bit windows, looks like a regression. Also loads fine on 1.16.0.dev0+d14632c

No issues on 1.18.0.dev0+84e8412 + python 3.5 either. Perhaps only the release has the problem, or perhaps it's a 3.7 issue

On 3.5 and 84e8412, that gives `ValueError: invalid shape in fixed-type tuple`. For some reason the pickle does not cause that error

My hunch is the cause is this line: https://github.com/numpy/numpy/blob/edf8a5f70bd946135a07bf8b1eada5feddef4b94/numpy/core/src/multiarray/descriptor.c#L501 which calls `print` while an exception is in flight. I have no idea why that print is there

Yep, that fixed it

Don't think we can close this, the problem I solved is the one you discovered, not the one this issue reports

Note that the result is usually a `datetime.timedelta`, if one is possible to create: ``` >>> np.array([10 ** 10], dtype='M8[us]').astype(object) array([datetime.datetime(1970, 1, 1, 2, 46, 40)], dtype=object) ```