I liked the idea of combining traceRunStatus and handleRunStatus, but
I think we lost a bit of clarity where traceRunStatus wants to fall
back to handleRunStatus when the breakpoint is enabled.  So I
refactored it a bit more.