Chris,

    I will look into this problem, but I am not sure how to reproduce it. Maybe you can provide me a test case? If so, please open a JIRA, attached your test case and ping me.

    Thanks,
        Edson
   

2010/8/10 Chris Raschl <c.raschl@dyless.com>
Hi,
Just in case someone faces the same problem, comments inline:

Am 05.08.2010 13:53, schrieb Chris Raschl:
> I'm using the latest drools and drools flow snapshot with jpa persistence.
>
> When I'm retracting a fact from the knowledge session (which causes
> logical retraction), the session can't be persisted any more.
>
> During retraction a LogicalRetractCallback is instantiated using this
> constructor:
>
> public LogicalRetractCallback(final TruthMaintenanceSystem tms,
>                                 final LogicalDependency node,
>                                 final Set set,
>                                 final InternalFactHandle handle,
>                                 final PropagationContext context,
>                                 final Activation activation) {
>       this.tms = tms;
>       this.node = node;
>       this.set = set;
>       this.handle = handle;
>       this.context = context;
> }
>
> As you can see an activation object can be passed (and actually is
> passed) to the constructor, but is never used. Retraction works though.
>
> The problem occurs when the knowledge session is persisted, because
> LogicalRetractCallback.write assumes that activation is not null, which
> leads to this npe:
>
> Caused by: java.lang.NullPointerException
> org.drools.common.TruthMaintenanceSystem$LogicalRetractCallback.write(TruthMaintenanceSystem.java:251)
> org.drools.marshalling.impl.OutputMarshaller.writeActionQueue(OutputMarshaller.java:201)

Adding this.activation = activation to the constructor does not help,
beause the npe in the write method remains. I tried it, then the whole
expression "context.terminalTupleMap.get( this.activation.getTuple() )"
in the write method evaluates to null...

I found out persistence was triggered by an unneccessary flush in my
code. I removed that, everything works now.

That seems to be the case because the action queue is empty then and
there's no need to persist it, thus the write method is never called. I
guess the problem remains if someone tries to persist a session with an
action queue containing a LogicalRetractCallback object.

- chris
_______________________________________________
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users



--
  Edson Tirelli
  JBoss Drools Core Development
  JBoss by Red Hat @ www.jboss.com