The previous topic decided when the run phase ends. This one covers everything around that moment: how a test turns counts into one trustworthy verdict, how phase domains and jumps change the schedule, why components sometimes run in the wrong order or not at all, and how to take apart a test that will not end. It finishes with the rarely justified tool of defining your own phase.
Name the four post-run phases, their traversal order, and the false-pass sources that slip past them.
Explain what a phase domain owns, and what sync and jump really do to the schedule.
Choose between a phase jump and a reset-aware run_phase for mid-test reset, and defend it.
Find a silent override, a late config set or a same-time-step race from the symptoms.
Triage a hang or an instant pass from the objection dump in minutes.
Lesson | What it answers |
|---|---|
Extract, check, report and final, one verdict, and the ways a test passes when it should not. | |
What a domain owns, how sync barriers join two schedules, and when a second domain pays off. | |
Mid-test reset two ways: jumping the schedule, or reset-aware components in run_phase. | |
Late overrides and config sets, methods that never run, and races inside one time step. | |
Hang, instant pass or early end: the dump, the trace and the fix for each. | |
Defining and inserting your own phase, and why reset-aware run_phase usually wins. |