Aleksey Shipilev wrote:
Good point, Nathan! It was DRLVM, but I'll try on IBM VME too.
FWIW the IBM VME available for Harmony doesn't do a good job of distributing the identityHashCode.
Anyway, I had incorporated this option in (4) question - can we just don't trust VM's System.identityHashCode and dope the hashcode with our own salt?
We had this discussion a awhile ago, and since we don't know what the distribution of the VM hashCode implementation will be, the only thing to do is assume that the VM is producing a good distribution, and let them suffer the consequences if they do not.
Regards, Tim
