Strange bug using ProcessPoolExecutor on Mac an Linux

[joelvbernier]

Was getting a weird error when trying to run a script that used a ProcessPoolExecutor with mp_context='fork' from 'spawn' processes. I tracked down at least one sneaky 'spawn' to the calculation of pixel_angles in hexrd.instrument. The fix, at least to start, is to force context to 'fork' where supported, and explicitly set the context in all instances of ProcessPoolExecutor.

[psavery]

Yeah, currently, Linux defaults to fork but Mac and Windows default to spawn.

I'm guessing this issue was on Mac? The multiprocessing docs indicate that spawn should be used for Mac. If that's the case, I think we should always be able to leave everything to their defaults?

Changed in version 3.8: On macOS, the spawn start method is now the default.
The fork start method should be considered unsafe as it can lead to crashes
of the subprocess. See **[bpo-33725](https://bugs.python.org/issue33725).**

https://docs.python.org/3/library/multiprocessing.html