dag.wan...@oracle.com (Dag H. Wanvik) writes: > Myrna van Lunteren <m.v.lunte...@gmail.com> writes: > >> I noticed this failing when I looked at the jdk 1.7 nightly tests a >> couple of days ago, and mentioned it in an email...I am more focused >> on 10.8.2 at the moment though. > > :) Very understandable. > >> >> Although it definitely popped up after that change to the test for >> DERBY-5044, I think you should log a separate issue. If it turns out >> to be a Java 7 release issue, it can be closed as Invalid later. But >> perhaps Mamta has a different opinion... > > Makes sense to me, can't harm, logged DERBY-5409.
It turned out to be some missing cleanup between test cases which only caused problems when the test cases ran in a particular order. Java 7 tended to pick that ordering more frequently than other JVMs. I've checked in a fix. -- Knut Anders
