This better implements the "negative event delay" configuration, where we actually want the background thread to get "ahead" of the foreground threads by making the foreground threads wait.
Happily, I havnen't observed any new test failures. (Test runs are here.)
This better implements the "negative event delay" configuration, where we actually want the background thread to get "ahead" of the foreground threads by making the foreground threads wait.
Happily, I havnen't observed any new test failures. (Test runs are here.)