Community
Participate
Working Groups
I20071120-1300 Test suites killed by timeout should appear red. Due to a bug on the Mac one of our tests hug and hence made the tests wait for 2hrs holding off the final results for everybody. Unfortunately we did not notice this so far because all was green. We should asap make killed test suites appear red. Idea: create a simple JUnit test that fails saying 'test suite $testSuiteName killed after timeout of $timeoutVal'. If a test suite is killed then run this dummy JUnit tests via ant task and make sure the output file has the expected name. I marked this as 'major' as this can cause the whole Eclipse test run to wait for 2hrs without being noticed.
Created attachment 83657 [details] Consolelog with timeout Here's the relevant extract of a consolelog.txt for a test that failed with time out.
Another alternative would be use an ant listener to implement this, I'm investigating.
Can we fix this for 3.5?
Ping. Kim, it would be good to get this fixed.
*** Bug 296060 has been marked as a duplicate of this bug. ***
Ping ;-)
I have to mark this as blocker since the consequences of this bug are very bad: I just found out that there's another test (suite) which gets killed in every build including 3.5. This means that this always contributes 2 hours to each test run without producing anything useful.
Is this really a blocker.. it's hard to see how a bug that is four years old blocks our development.
(In reply to comment #8) > Is this really a blocker.. it's hard to see how a bug that is four years old > blocks our development. Well, it blocked the results from appearing for 2 hours for every build over more than a year. The bad part is that this can happen again at any time and we won't notice it.
Looks like this has been deferred several times already ... and, I will once again. But, will look to improve in Luna ... if nothing else, distinguish the different types of "DNFs" (for which there is another bug).
Interesting observation from bug 417928 / bug 418101: The test suite should appear read if - the <java> Ant task killed the test run OR - org.eclipse.test.EclipseTestRunner#startStackDumpTimoutTimer() calls dump(0) In the bug 418101 scenario, the call to dump(0) causes a call to display#syncExec(..), which wakes the Display from its 2h-epsilon sleep. The 5s before the second call to dump(..) are enough to make the whole test suite finish and appear in a shiny green. In the end, the test is NOT killed by the Ant task timeout, but turned green by the observer -- a perfect heisenbug.
(In reply to Markus Keller from comment #11) > Interesting observation from bug 417928 / bug 418101: > > The test suite should appear read if > - the <java> Ant task killed the test run OR > - org.eclipse.test.EclipseTestRunner#startStackDumpTimoutTimer() calls > dump(0) > Thanks for finding and documenting this issue ... I was content to leave it as a mystery. :) I fear that the first condition above ("Ant task killed the test run") is currently simply "inferred" by the absence of test results (which is what leads to bug 399636) so my guess is we need to stop relying on "absence of results" for anything, and simply learn how to scan the log(s) for "timed out" (and other, serious) error messages.
I think we really have to perform both checks. The "absence of results" could also happen for other reasons, e.g. VM crash, System.exit(), hardware failure, ... > simply learn how to scan the > log(s) for "timed out" (and other, serious) error messages. Probably a better solution would be to write out a special file <testname>.timeout.txt to the consolelogs directory. That file could also contain the stacktraces, making them easier to find.
Nothing planned for Luna SR0, so untargettng.
Re targetting to 4.8
There are plans for this. Will revisit in 4.10
Moving out of 4.10. Please re triage appropriately
This bug hasn't had any activity in quite some time. Maybe the problem got resolved, was a duplicate of something else, or became less pressing for some reason - or maybe it's still relevant but just hasn't been looked at yet. As such, we're closing this bug. If you have further information on the current state of the bug, please add it and reopen this bug. The information can be, for example, that the problem still occurs, that you still want the feature, that more information is needed, or that the bug is (for whatever reason) no longer relevant. -- The automated Eclipse Genie.
