Logic of nova-cells test case 'test_cells_scheduler.test_filter_schedule_skipping' is odd

[Wenzhi Yu]

Logic of nova-cells test case 'nova.tests.unit.cells.test_cells_scheduler.test_filter_schedule_skipping'[1] is really odd:

def test_filter_schedule_skipping(self):
 # if a filter handles scheduling, short circuit

 def _grab(filter_properties):
  return None

 self.stubs.Set(self.scheduler, '_grab_target_cells', _grab)

 def _test(self, *args):
  raise test.TestingException("shouldn't be called")

 try:
  self.scheduler._schedule_build_to_cells(None, None, None, _test,
            None)
 except test.TestingException:
  self.fail("Scheduling did not properly short circuit")

It uses a try-except block to ensure that '_test()' function will not be called during case execution, which is odd and poorly readable.

[1]https://github.com/openstack/nova/blob/13.0.0.0rc1/nova/tests/unit/cells/test_cells_scheduler.py#L354-L369

[Wenzhi Yu (yuywz)]

Gerrit code review: https://review.openstack.org/#/c/285146/

[Sylvain Bauza (sylvain-bauza)]

I don't feel it's a bug. You should just provide a change, without really needing to add this one.

[Markus Zoeller (markus_z) (mzoeller)]

In addition to comment #2:
The commit message of the change is documentation enough. No need to go through the bug triage process.

[Wenzhi Yu (yuywz)]

Gerrit code review: https://review.openstack.org/#/c/285146/