[Bug]: Intermittent test failures with tests using time-machine #15942
Description
Is there an existing issue for this?
- I have searched the existing issues
What happened?
Some of our tests are intermittently failing since the switch from freezegun to time-machine (#15892).
It looks like we have those failures when we use travel(datetime.now()) and then check the timestamp, because of microseconds. test_process_event_email_change() is one we've seen fail repeatedly (approximately 1 in 10 times), and the following modification makes it fail consistently on my machine:
@mock.patch('olympia.accounts.views.primary_email_change_event.delay')
def test_process_event_email_change(self, event_mock):
now = datetime.now().replace(microsecond=975982)
with time_machine.travel(now.timestamp(), tick=False):
FxaNotificationView().process_event(
self.FXA_ID,
FxaNotificationView.FXA_PROFILE_CHANGE_EVENT,
{'email': 'new-email@example.com'},
)
event_mock.assert_called_with(
self.FXA_ID, now.timestamp(), 'new-email@example.com'
)Fails with:
/usr/local/lib/python3.12/unittest/mock.py:949: in assert_called_with
raise AssertionError(_error_message()) from cause
E AssertionError: expected call not found.
E Expected: delay('ABCDEF012345689', 1764162587.975982, 'new-email@example.com')
E Actual: delay('ABCDEF012345689', 1764162587.9759817, 'new-email@example.com')
What did you expect to happen?
No intermittent failures
┆Issue is synchronized with this Jira Task