Added @pitrou as a reviewer for the PR, since he's an active expert for the multiprocessing module.

Thanks for the PR @awecx.

I confirmed from the source code that AsyncResult.successful() does indeed raise a ValueError instead of an AssertionError:

cpython/Lib/multiprocessing/pool.py

Lines 753 to 756 in a796d8e

	def successful(self):
	if not self.ready():
	raise ValueError("{0!r} not ready".format(self))
	return self._success

LGTM. This is also present in the 3.8 and 3.7 branches, so this PR should be backported to those versions as well.

Thanks @aeros for the clarification. Indeed I should have proven the typo claim more clearly:
@pitrou if double check is necessary, you can run the following snippet:

from multiprocessing import Pool
from time import sleep

with Pool(processes=1) as pool:
    result = pool.apply_async(sleep, (10000,))
    result.successful()