Unless the test cases are incomplete. Or the CI jobs are configured to run tests independently and off-cadence so that code that would break is tested after a m

Unless the test cases are incomplete. Or the CI jobs are configured to run tests independently and off-cadence so that code that would break is tested after a merge. Or the tests have a bug in them. And so on.