| Re: [qvto-dev] Strange Completion Testing results |
Hi Ed, +1 in bugzilla [1].I can't reproduce the failure in local, but the test cases passes with the fix, so go ahead with it.
[1] https://bugs.eclipse.org/bugs/show_bug.cgi?id=402809 Cheers, Adolfo. On 09/03/2013 12:20, Ed Willink wrote:
Hi I've just eliminated the org.junit4 dependency, since M6 will no longer provide this JUnit 3/4 transition support. [1] Yesterday this was a complete disaster and the org.eclipse.m2m.internal.tests.qvt.oml.ui.AllTests failed 100% and repeatedly so on Hudson. I started trying to debug the problem and [2] provides a patch that at least avoids the NPE that makes these tests rather temperamental. Today everything works perfectly. So my best guess is that some bug has been fixed in the platform overnight. Slave1 in use for both successes and failures. Anyway, the patch in [2] may be useful to anyone who has a reproducible way of getting org.eclipse.m2m.internal.tests.qvt.oml.ui.AllTests failures. Please try it. Regards Ed Willink [1] https://bugs.eclipse.org/bugs/show_bug.cgi?id=402746 [2] https://bugs.eclipse.org/bugs/show_bug.cgi?id=402809 _______________________________________________ qvto-dev mailing list qvto-dev@xxxxxxxxxxx http://dev.eclipse.org/mailman/listinfo/qvto-dev