In message <201005261641.o4qgf7hx023...@d06av03.portsmouth.uk.ibm.com>, Mark Hindess writes: > > In message <4bfb875a.6000...@googlemail.com>, Oliver Deakin writes: > > > > On 25/05/2010 09:13, Tim Ellison wrote: > > > I tested on Windows XP, MS VS 2003, Ant 1.7.0 > > > > > > Same results as 5.0 branch I reported yesterday, with additional > > > failures in JDWP tests: > > > > > > ClassPrepareTest#testClassPrepareEventWithSourceDebugExtension001 > > > ClassPrepareTest#testClassPrepareEventWithSourceDebugExtension002 > > > ClassPrepareTest#testClassPrepareEventWithSourceDebugExtension003 > > > > > > All failed in JDWPTestCase.java line 752 with: > > > Invalid event kind, expected:<8(CLASS_PREPARE)> but was:<99(VM_DEATH)> > > > > I see a similar set of failures in my first test run onWindows x86. I'm > > rerunning the 6.0 tests at the moment to check if my failures are > > consistent. Apart from those 3 I did not see any other unusual failures. > > I see these failures if I run them together but I can't reproduce them > if I try to debug the test with: > > target/hdk/jdk/jre/bin/java \ > -cp target/hdk/build/test/jpda/jpda_tests.jar:target/hdk/build/test/juni > t.jar org.junit.runner.JUnitCore \ > org.apache.harmony.jpda.tests.jdwp.MultiSession.ClassPrepareTest > > We need to take a better look at these jdwp issues after the releases.
Actually it says it is passing but actually it is reporting: STDERR> java.lang.ClassNotFoundException: org.apache.harmony.jpda.tests.jdwp.Events.SourceDebugExtensionMockClass STDERR> at java.lang.Class.forName(Class.java:129) STDERR> at org.apache.harmony.jpda.tests.jdwp.Events.ClassPrepareDebuggee.run(ClassPrepareDebuggee.java:55) STDERR> at org.apache.harmony.jpda.tests.share.Debuggee.runDebuggee(Debuggee.java:84) STDERR> at org.apache.harmony.jpda.tests.jdwp.Events.ClassPrepareDebuggee.main(ClassPrepareDebuggee.java:38) because the jdktools/modules/jpda/src/test/resources is missing from the jdpa_tests.jar. Adding this path to the -cp makes it pass. I'll fix this after release. This might cause the VM_DEATH when run as a batch I suppose ... but since we weren't previously including any of the jpda tests in the releases it is not exactly a regression. Regards, Mark.
