It would be good to figure this out, but that probably involves rewriting some of the tests to run in child processes where the asynchronous context isn't being affected by whatever changes have been made to tap that cause the asynchronous event count to be greater than it was when the tests were written.
It would be good to figure this out, but that probably involves rewriting some of the tests to run in child processes where the asynchronous context isn't being affected by whatever changes have been made to tap that cause the asynchronous event count to be greater than it was when the tests were written.