Interesting issue. I wonder if we are hitting this bug http://bugs.sun.com/view_bug.do?bug_id=5091934. The intent of the test is to make sure all 10 threads execute serially. Since thread.sleep is not guaranteed to be exact on Windows, we need to change the test expectation. Dropping it to 1240 might work for that.
Any other thoughts? On 8/22/12 7:02 AM, "Lasschuit, JIJ (Hans)" <[email protected]> wrote: >Hi everybody, > >When doing a rave mvn install the testThreadWithSharedKeys in >DefaultLockServiceTest.jave failed. It gets a value of 1243L while >expecting a value of 1250L or greater. Could this be due to the speed of >the computer it is running on (Windows server 2008R2 64-bit on VMWARE >running on 2 Intel Xeon x5690 processors at 3.47 Ghz an 4 GB memory)? >Would it be safe to adjust the testvalue to a asserted value of e.g. >1240L to make the install process succeed the test? > >Thanks and best regards, >Hans
