Community
Participate
Working Groups
It seems that the test ResourceComputationSchedulerWithEventBusTest.testDemandShutdownWithLongRunningTaskThatInterruptsGracefully is flaky. I had a test failure in master-build-on-egit-logical-nightly #551, retriggered it, and #552 went through. The failure trace was as follows: java.lang.AssertionError: null at org.junit.Assert.fail(Assert.java:86) at org.junit.Assert.assertTrue(Assert.java:41) at org.junit.Assert.assertTrue(Assert.java:52) at org.eclipse.emf.compare.ide.ui.tests.logical.resolver.ResourceComputationSchedulerWithEventBusTest.testDemandShutdownWithLongRunningTaskThatInterruptsGracefully(ResourceComputationSchedulerWithEventBusTest.java:280)
I don't think I've seen this failure yet, has it been reproduced somehow?
I have never succeeded to reproduce it locally, but it indeed happened very recently on the master-gerrit build: https://hudson.eclipse.org/emfcompare/job/master-gerrit/3352/PLATFORM=mars,jdk=JDK_1.8.0_Latest/
I was just wiping the workspace for the master-build-on-gerrit-logical-nightly to test :) So this indeed still happens, looking at the code though, I'm not sure how we can ensure it won't... or if it's somehow revealing an underlying bug-in-waiting in our code
New Gerrit change created: https://git.eclipse.org/r/104665
I'll increase the scheduler's timeout from 10ms to 100ms to see if this solves things.
Gerrit change https://git.eclipse.org/r/104665 was merged to [master]. Commit: http://git.eclipse.org/c/emfcompare/org.eclipse.emf.compare.git/commit/?id=591ce92e5689695a3ae3a5d49db6fc0ff93d5c33
Closing for now, will reopen if we reproduce again
Reproduced again with https://hudson.eclipse.org/emfcompare/job/master-gerrit/3352/ We need to debug through this test and find a way to ensure the scheduled computation is properly interrupted before the scheduler times it out.