After doing some digging it appears that the errors and other failed tests related to the kit registry were a red herring. What it appears was happening was that on a slow CI runner if one of the tests with chained calls to dispatch_after would time out, it would move on to the next tests but then a few seconds later it would finish which interfered with the data of other tests. The timing was set pretty tight, or rather plenty of time on a fast MacBook Pro, but tight on a slow CI runner. I've increased all timeouts to 15 seconds, which should be more than enough time for all runners to complete.
Testing Plan
[x] Was this tested locally? If not, explain why.
Tests ran fine locally, but they always did... I'll be running them multiple times via this PR to confirm it succeeds every time.
Reference Issue (For mParticle employees only. Ignore if you are an outside contributor)
Summary
After doing some digging it appears that the errors and other failed tests related to the kit registry were a red herring. What it appears was happening was that on a slow CI runner if one of the tests with chained calls to dispatch_after would time out, it would move on to the next tests but then a few seconds later it would finish which interfered with the data of other tests. The timing was set pretty tight, or rather plenty of time on a fast MacBook Pro, but tight on a slow CI runner. I've increased all timeouts to 15 seconds, which should be more than enough time for all runners to complete.
Testing Plan
Tests ran fine locally, but they always did... I'll be running them multiple times via this PR to confirm it succeeds every time.
Reference Issue (For mParticle employees only. Ignore if you are an outside contributor)