I'm testing on JDK7 on a mac right now. When I see a good test run (started one last night. Looks promising at the moment) I was going to fire up tests on JDK6/Solaris 10, JDK7/win7x64 and JDK7/winxp32.
Given that the Levels fix is for a JDK7 issue, it's pretty critical that we test on JDK7. Even having said that, I don't know that the Levels issue is immediately apparent - we should probably write a test that demonstrates the issue and fix. Is Dennis there? Do you have some code that shows the Levels problem? If so I'll see about writing a jtreg test for it. Cheers, Greg. On Wed, 2013-04-10 at 07:12, Peter Firmstone wrote: > Greg, > > Are you testing on Java 7? > > The last release was only tested with Java 5 and 6. > > Can you test with Java 6? > > Regards, > > Peter. > > > On 10/04/2013 6:32 AM, Greg Trasuk wrote: > > On Tue, 2013-04-09 at 15:29, Peter Firmstone wrote: > >> This was being discussed in two separate threads: > >> > >> Greg, > >> > >> Are you able to find which process is using a port? Perhaps jvisualvm > >> might help determine which test it is? If you can narrow it down to a > >> bunch of test definitions sharing a test class, by identifying the class > >> on the stack, you'll find it. > >> > > No, there's no process using the port. The evidence points to a > > configuration error. The problem is not that the registrars don't > > start. It's that they start on random ports when the port is supposed > > to be specified.
