This commit adds an opt-in scheduler option for dynamic scheduling.
Instead of partitioning the test list up-front based on historical
timing data this commit lets each worker ask for the next test
dynamically.  This is built using python's multiprocess module to
launch new workers instead of shelling out to call python via
subprocess.

This hopefully will provide a better worker balance since we will keep
each worker occupied until there are no more tests to be run. Instead
of trying to pack fill each work optimially up front. Additionally this
should hopefully improve the pdb story for users who use pdb with tests.
Since instead of spawning subprocesses calling python to invoke the
subunit runner and reading the subunit stream from stdout and instead
uses multiprocessing to fork workers and uses pipes to pass the subunit
streams between workers.

Co-Authored-By: Adam Spiers <github@adamspiers.org>

This commit fixes the failing tests by catching a couple of missing
things from the update. The biggest fix was that for the --no-discover
case we still use a subprocess and because of that we need to tell
output.ReturnCodeToSubunit to that the input is not dynamic (and
therefore a Popen object) so it can handle that properly. The other
major change is that the return code tests are updated so that the
stdout and stderr from the subprocess calls are always decoded in the
non-subunit test cases. This was done primarily for ease of debugging,
but it also enabled the removal of several decode() calls when the
output is parsed.

This is a refinement on the previous commit to reduce unecessary changes
to the functional tests in the test_return_codes module. Mainly always
decoding the output from the subprocess for testing broken things
unexpectedly when a bytes object was expected.

Conflicts:
	stestr/commands/run.py
	stestr/output.py
	stestr/test_processor.py
	stestr/tests/test_return_codes.py

I originally developed this feature when we still supported older
python versions in stestr. The dynamic scheduling feature depends on
functionality added in Python 3.5. Since then the WIP feature branch sat
stale for years since that time we've bumped the minimum version of
Python supported to 3.7 so the runtime check for older python versions
is no longer needed.