[Bug general/4996] New: NullPointerException in testSingleExistingCloneAttached causes next test to hang

[Bug general/4996] New: NullPointerException in testSingleExistingCloneAttached causes next test to hang

[kris dot van dot hees at oracle dot com]

This problem has been seen both on x86 and x86_64.  The TestRunner output shows:

Running testMultiThreadedStoppedAckDaemon(frysk.proc.TestProcStopped) ...PASS
Running testMultiThreadedStoppedDetached(frysk.proc.TestProcStopped) ...PASS
Running testMultiThreadedStoppedAttached(frysk.proc.TestProcStopped) ...PASS
Running testMultiThreadedRunningAckDaemon(frysk.proc.TestProcStopped) ...PASS
Running testMultiThreadedRunningDetached(frysk.proc.TestProcStopped) ...PASS
Running testMultiThreadedRunningAttached(frysk.proc.TestProcStopped) ...PASS
Running testManyExistingThreadAttached(frysk.proc.TestProcTasksObserver) ...PASS
Running testSingleExistingCloneAttached(frysk.proc.TestProcTasksObserver) ...ERROR
  java.lang.NullPointerException
Running testCloneThenKillAttached(frysk.proc.TestProcTasksObserver) ...

At this point the testsuite execution is hanging, with the following child
process of TestRunner sticking around:

frysk-core/frysk/pkglibdir/funit-slave -t 5 -m <pid>

Killing the child process manually from a command shell helps TestRunner continue.

-- 
           Summary: NullPointerException in testSingleExistingCloneAttached
                    causes next test to hang
           Product: frysk
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: general
        AssignedTo: frysk-bugzilla at sourceware dot org
        ReportedBy: kris dot van dot hees at oracle dot com


http://sourceware.org/bugzilla/show_bug.cgi?id=4996

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug general/4996] NullPointerException in test case using funit-slave causes next test to hang in waitAll

[kris dot van dot hees at oracle dot com]

------- Additional Comments From kris dot van dot hees at oracle dot com  2007-09-06 20:47 -------
Changed the summary due to new evidence collected in a manual invocation of
TestRunner -c FINE.  More details in the next update...

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|NullPointerException in     |NullPointerException in test
                   |testSingleExistingCloneAttac|case using funit-slave
                   |hed causes next test to hang|causes next test to hang in
                   |                            |waitAll


http://sourceware.org/bugzilla/show_bug.cgi?id=4996

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug general/4996] NullPointerException in test case using funit-slave causes next test to hang in waitAll

[kris dot van dot hees at oracle dot com]

------- Additional Comments From kris dot van dot hees at oracle dot com  2007-09-06 20:54 -------
Created an attachment (id=1985)
 --> (http://sourceware.org/bugzilla/attachment.cgi?id=1985&action=view)
TestRunner -c FINE output


-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=4996

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug general/4996] NullPointerException in test case using funit-slave causes next test to hang in waitAll

[kris dot van dot hees at oracle dot com]

------- Additional Comments From kris dot van dot hees at oracle dot com  2007-09-06 20:55 -------
Attached is TestRunner -c FINE output illustrating the problem.  Note that the
first listed test fails to complete the process of creating a child and using
it.  The child process *does* get created on the OS side, but the frysk core
code fails to obtain the Task Id (due to NullPointerException getting thrown -
see bug #4999).  The child process exists, but never gets added to the list of
processes to tear down at the end of the test.

The next test execution also uses a child (successfully) and tears it down at
the end.  However, the waitAll operation in the tear down process seems to get
stuck due to the stray child from the previous test.

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=4996

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug general/4996] NullPointerException in test case using funit-slave causes next test to hang in waitAll

[kris dot van dot hees at oracle dot com]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|frysk-bugzilla at sourceware|kris dot van dot hees at
                   |dot org                     |oracle dot com
             Status|NEW                         |ASSIGNED


http://sourceware.org/bugzilla/show_bug.cgi?id=4996

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.