Looks like a problem with the IcedTea jdk when retrieving the
AccessControlContext. River appears to have handled the socket closure
properly.
The stress test doesn't stress River much, it uses TaskManager which causes
contention within the test code. TaskManager is only retained for
Is this an actual problem with River 3.0, or just an indication that my
Linux box is not tough enough for this stress test?
[java] Running
org/apache/river/test/impl/outrigger/matching/StressTestInterleavedWithShutdown.td
[java] Time is Sat Sep 17 12:36:29 PDT 2016
[java]