Thanks for explaining this. You are right. There is actually no issue
involved here. It was purely my misinterpretation of the output.
Actually I did not factor in the fact that a thread switching (from the
test thread to the active-object intrinsic thread) could happen before
the activation list is displayed in the test thread. By the time the
test tread is re-scheduled, the intrinsic thread may have already
processed and cleared up the activation list.