CPython's multiprocessing package is implemented fairly differently on Windows and on Linux, as a Windows implementation cannot rely on fork(2). However, it seems to me that the Windows implementation of multiprocessing (spawning a separate process and send it the required state by serializing it) should work on Linux (or am I wrong?).
While I work on Linux, I would like to make sure that the code I write also works on Windows (e.g., not accidentally have unpicklable arguments, etc.). Is there a way I can force CPython to use the Windows implementation of multiprocessing on Linux?
Thanks.