Closed GoogleCodeExporter closed 8 years ago
Interesting. I believe the problem is that datanucleus is attempting to add
the new
B instance to the L2 cache before its internal object id has been assigned. As
a
result, a "temporary" internal object id is constructed, which is really just a
wrapper around the object itself that implements hashCode() using
System.identityHashCode() and equals() using reference compare. This wrapper
is not
Serializable, so memcache rejects it.
Original comment by max.r...@gmail.com
on 27 Nov 2009 at 11:23
Even with NUCCORE-540 "fixed", this issue is still around -- see additional
comments:
* http://www.datanucleus.org/servlet/jira/browse/NUCCORE-540
Original comment by ales.justin
on 17 Aug 2010 at 2:24
"The issue is still around". Well no testcase was ever provided for that, so
the onus is on the user to demonstrate that the issue is around in DataNucleus
code. If you provide one then it will be looked at.
Original comment by googleco...@yahoo.co.uk
on 18 Jul 2011 at 4:52
Not reproduceable currently, so downgrading
Original comment by googleco...@yahoo.co.uk
on 21 Sep 2011 at 3:59
Can't reproduce (with SVN trunk, and that uses DN 3.0 that has a completely
rewritten L2 cache). Provide a complete testcase and it could be moved forward
Original comment by googleco...@yahoo.co.uk
on 1 Nov 2011 at 4:25
Original issue reported on code.google.com by
andr.m...@gmail.com
on 21 Nov 2009 at 3:01