Warren Weckesser

The script to find the missing imports was missing the `signal.windows` submodule. The script and its output above have been updated.

Another small update: the script was not checking `scipy.datasets`, so it missed `electrocardiogram`. The script and the output shown above have been updated.

Thanks @BassCoder2808, but in fact I am about to submit a pull request to finish these off. The main purpose of the pull request will be to fix a bunch...

Another update: the functions in `scipy.fft` and `scipy.fftpack` that are listed above are actually NumPy functions, so they can be skipped.

https://xkcd.com/1296/ 😁

@andyfaff wrote > Some docstrings have multiple examples, but these are often interspersed with lots of text. Should a numpy import happen once in a docstring, or with each example...

I'm +0. The arguments for not including the import have never seemed very compelling, but issues like this one have rarely come up, so I guess it is rare for...

This could also be a useful feature for the univariate distributions, where the "hyperrectangle" is just an interval. It is common to want the probability of an interval `a

`float16` is being used "in the wild". I ran into this issue while investigating http://stackoverflow.com/questions/44003497/large-memory-usage-of-scipy-sparse-csr-matrix-toarray. I guess we should check at construction to avoid surprises in later operations.

@larsoner wrote > you should be able to remove the pytest.mark.timeout mark Done. I left the code for handling custom marks in `test_moments`, with a comment about how it might...