Learn more about these different git repos.
Other Git URLs
I found here an example of a message from greenwave:
https://apps.fedoraproject.org/datagrepper/id?id=2017-db45b07c-b39a-4fd3-af49-ae7cb83bb78b&is_raw=true&size=extra-large
It looks like greenwave decided to publish because the details of the failure changed. It used to be that 2 of 3 tests were failing. Now 1 of 3 tests are failing. However, the overall decision is still the same: policies_satisfied: false.
policies_satisfied: false
We should only publish messages when the satisfaction value changes.
Metadata Update from @ralph: - Issue assigned to ralph
I guess I don't understand why this is happening, given the existence of test_no_message_for_unchanged_decision in our test suite.
test_no_message_for_unchanged_decision
OH! It may be related to #103... since we're sending a message for each policy, that can be a problem.
See #105.
I was actually thinking what problem would be caused, but I can't think of any except that there could be duplicate messages sent if there are two policies using same decision context, product version and test casename. I guess it would rarely happen to have policies like that. But, yeah, I agree that we should publish a message for every decision context and product-version combination which is a lot clearer.
Metadata Update from @ralph: - Issue status updated to: Closed (was: Open)
Metadata Update from @dcallagh: - Issue close_status updated to: Rejected
Login to comment on this ticket.