Thanks Richard: That was a great help: http://stackoverflow.com/questions/32871715/why-newglobalref-will-core-dump-on-jni-on-an-ibm-jdk/32905153#32905153
This confirms the fix I made as valid: https://github.com/apache/activemq-artemis/commit/18e41963b61a5ef6dbcdb150e27c07c062deaa66 And I helped a lot through this issue.. thanks: https://github.com/apache/activemq-artemis/commit/c89316623f4abf68911699160d24747898d7dd77 On Wed, Sep 30, 2015 at 4:42 AM, Richard Kettelerij < [email protected]> wrote: > Hi Clebert, > > As far as I know IBM doesn't have a public issue tracker for JVM issues. > But I do know someone from the IBM JVM team reached out to the guys from > ElasticSearch since they were (also) having some trouble with the IBM JVM. > See > > https://discuss.elastic.co/t/need-help-with-ibm-jdk-issues-with-es-1-4-5/1748/5 > for a reply by George Chan from the IBM JVM team. He suggests posting > issues to https://developer.ibm.com/answers/questions with hashtag > #IBMJDK. > Maybe you can do the same or try to contact him directly? > > Regards, > Richard > > On Wed, Sep 30, 2015 at 2:49 AM, Clebert Suconic < > [email protected]> > wrote: > > > Anyone knows a place to raise issues on the IBM JDK? > > > > > > The IBM JDK will core dump if you create a global Ref for MethodIDs on > the > > JNI. This behaves fine on the Sun JDK and will cause a GPF / core dumped > on > > a IBM JDK. > > > > > > Even thought I didn't apparently need the Gobal Ref on the Method ID, it > > would be nice to raise this issue there. > > > > > > I'm fixing this on the Artemis Native code.. and it would be nice to > raise > > this as an issue on the IBM JDK. > > > -- Clebert Suconic
