a small issue with the global_variable+pool approach when using Python 3.8 in macOS

I just upgraded to Python 3.8 on a Mac and found that my previously working code (under 3.7) broke down, as well as the "Multiprocessing+global-variable" method presented at the end of this tutorial. It turns out Python 3.8 in macOS switches to the spawn start method in multiprocessing as default, so the resources of the parent process may not be inherited by the child process. Then the "global" variables become inaccessible.

A short workaround may be changing

from multiprocessing import Pool

into

import multiprocessing as mp
Pool = mp.get_context('fork').Pool

in the tutorial example. This is just to make sure that the fork method is always used.

This issue is not an emcee problem at all, and will only happen for certain Python versions + platforms. But the tutorial could be slightly tweaked for this.

btw, thank you for writing the excellent documentation for the package.

Start with the parallel tutorial linked in the issue, focusing on the global-variable and Pool example and its multiprocessing imports. Check the Python 3.8 macOS spawn behavior described there, then update the example so its documented workaround consistently uses fork; the tutorial should explain how the example remains usable on that platform.