On at least x86-64 RHEL. At a technical level, this makes little sense, frysk should know about all children suggesting that waitpid event is stray. 3) testBlockingFibonacciFork(frysk.proc.TestTaskObserverBlocked)java.lang.RuntimeException: Unknown waitpid stopped event (process 5342) at frysk.sys.Wait.waitAllNoHang(funit) at frysk.proc.LinuxPtraceHost$PollWaitOnSigChld.execute(funit) at frysk.event.EventLoop.runEventLoop(funit) at frysk.event.EventLoop.runPolling(funit) at frysk.proc.TestLib.assertRunUntilStop(funit) at frysk.proc.TestLib.assertRunUntilStop(funit) at frysk.proc.TestTaskObserverBlocked$BlockingFibonacci.<init>(funit) at frysk.proc.TestTaskObserverBlocked$5$ForkFibonacci.<init>(funit) at frysk.proc.TestTaskObserverBlocked.testBlockingFibonacciFork(funit) at frysk.junit.Runner.runCases(funit) at frysk.junit.Runner.runArchCases(funit) at frysk.junit.Runner.runTestCases(funit) at funit.main(funit)
RHEL: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=225449
This test is currently (x86_64 fedora 8) giving me: testBlockingFibonacciClone(frysk.proc.live.TestTaskObserverBlocked)java.lang.RuntimeException: {frysk.proc.live.LinuxPtraceTask@30cddf55,pid=7820,tid=7821,state=detached} in state "detached" did not handle handleStoppedEvent SIGSTOP(19)
Test passes always (./TestRunner -r 100 frysk.proc.live.TestTaskObserverBlocked) on both 2.6.23.1-42.fc8 x86 and 2.6.24.3-34.fc8 x86_64.