[Mono-bugs] [Bug 72673][Cri] Changed - NUnit threading tests fail on 2.0

bugzilla-daemon@bugzilla.ximian.com bugzilla-daemon@bugzilla.ximian.com
Mon, 18 Apr 2005 08:39:05 -0400 (EDT)

Please do not reply to this email- if you want to comment on the bug, go to the
URL shown below and enter your comments there.

Changed by flashdict@gmail.com.


--- shadow/72673	2005-02-16 22:26:39.000000000 -0500
+++ shadow/72673.tmp.11394	2005-04-18 08:39:05.000000000 -0400
@@ -1,20 +1,85 @@
 Bug#: 72673
 Product: Mono: Runtime
 Version: 1.1
+OS: unknown
 OS Details: 
 Status: NEW   
+Severity: Unknown
 Priority: Critical
 Component: io-layer
 AssignedTo: dick@ximian.com                            
 ReportedBy: bmaurer@users.sf.net               
 QAContact: mono-bugs@ximian.com
 TargetMilestone: ---
 Summary: NUnit threading tests fail on 2.0
 In r40784, I disabled threading tests because they are fucking our 2.0 test
 runs (many of them seem to hang). None of this happens on 1.0.
+------- Additional Comments From flashdict@gmail.com  2005-04-18 08:39 -------
+This problem appears to be isolated to the nunit tests in
+System.Runtime.Remoting.  At least on my own machine, nunit testing
+only hangs there.
+Running the net_2_0 profile, I get intermittent hangs about 50% of the
+time.  The rest of time, all of the tests complete successfully, but
+always reporting the following error:
+** (../../class/lib/net_2_0/nunit-console.exe:82786): WARNING **:
+Aborting of threads in domain
+domain-System.Runtime.Remoting_test_net_2_0.dll timed out.
+This message may be why this bug is suspected to be against io-layer.
+ So... turning on debugging in handles.c and threads.c, the test suite
+gets to the third test in the list, which is:
+***** MonoTests.Remoting.ContextsTest.TestRemoteContext
+This test allocates a new handle of type Event:
+** Message: _wapi_handle_new: Creating new handle of type Event
+** Message: _wapi_handle_new: Allocated new handle 0x23dc
+The tests progress, but this handle is never freed until the test:
+***** MonoTests.Remoting.HttpSyncCallTest.TestInterfaceComplexParamsInOut
+After not less than 180 attempts are made to free the handle.
+The tests then continue to progress up until
+***** MonoTests.Remoting.HttpSyncCallTest.TestInterfaceProcessContextData
+This test creates another event handle,
+** Message: _wapi_handle_new: Creating new handle of type Event
+** Message: _wapi_handle_new: Allocated new handle 0x24ec
+We then proceed to the Async test:
+***** MonoTests.Remoting.HttpAsyncCallTest.TestInstanceSimple
+And this test also spins for a terribly long time, waiting for the
+Event handle to be freed.
+Finally, the test:
+***** MonoTests.Remoting.HttpAsyncCallTest.TestAbstractSimple
+is run, and this is the first place we see a real hang.  Several
+thread, socket, and event handles are created and destroyed, then the
+wapi code gets into _wapi_handle_update_refs(), and never comes out
+again.  I started the test more than two hours ago, and it's still
+executing the TestAbstractSimple test, stuck in the
+_wapi_handle_update_refs() loop, with more than 150 executions (with
+8sec intervals between each one) so far.
+You can find the logfile which documents the above at:
+Hope this helps a bit.