While iterating on a unit test in db_s_config_server_test I started getting a new, unexpected error

ShardingTestFixtureCommon::tearDown() must have been called before destruction

This was surprising because I hadn't touched anything about the setup/teardown of the test. After some investigation, ultimately disabling the invariant in question, I saw a new exception that my disk had run out of space (seems to be due to something like SERVER-38689 but for sharding tests). I think what's happening here is that the destructor of ShardingTestFixtureCommon is swallowing top-level exceptions since its throwing its own when the invariant fails. It would be great to just report the top-level exception, since it's the real problem.