[rules-users] NPE on retract
Edson Tirelli
tirelli at post.com
Wed Apr 25 09:55:21 EDT 2007
Chris,
> My misunderstanding was that Drools used object identity with an
IdentifyHashMap unless it was configured to use a regular HashMap. Is this
not the case?
Yes, it uses identity as default for regular assertions, but at the same
time, it uses equality for the Truth Management System (TMS). As the TMS is
always on in the 3.0.x versions, the equalities are being verified even if
you don't use logical assertions. In the new version, it is possible to
explicitly turn off the TMS.
> Concerning not having equals/hashcode changing while the object is
asserted into working memory, is it ok for them to change as long as it is
followed by a "modify("? Is there different behavior between calling
"modify" and calling "retract and re-assert"?
In 3.0.x we have a "true modify" that means that a modify is not
retract+assert. This way we can avoid shadow facts and yet be consistent, as
long as the object "identity" is not changed. So, changing object identity
in 3.0.x may eventually cause unexpected behavior, even if you call modify()
later. In the new version, we added shadow fact support, so we can work with
object "identity" changes, but again, conceptually is not really good to
allow "fact mutation" :), so, you may want to avoid that anyway.
[]s
Edson
2007/4/25, Chris West <crayzfishr at gmail.com>:
>
> I'll try to create a self contained test. It may be a difficult task. The
> error seems to come and go depending on which other rules I leave in my rule
> set. In particular, there is another rule that has many conditions and only
> logs a message as the consequence that when it is included in the rule set,
> this error occurs. That particular rule has nothing to do with the type of
> object being retracted.
>
> My misunderstanding was that Drools used object identity with an
> IdentifyHashMap unless it was configured to use a regular HashMap. Is this
> not the case?
>
> Concerning not having equals/hashcode changing while the object is
> asserted into working memory, is it ok for them to change as long as it is
> followed by a "modify("? Is there different behavior between calling
> "modify" and calling "retract and re-assert"?
>
> Thanks for you help,
> -Chris
>
> On 4/24/07, Edson Tirelli <tirelli at post.com> wrote:
> >
> >
> > Chris
> >
> > Can you provide a self contained test that shows the problem and open
> > a JIRA with it?
> > I will take a look in it.
> >
> > There is no requirement to define equals/hashcode, but if you do so,
> > you need to make them time consistent, i.e., they cannot change during
> > the time the object is asserted into the working memory or you may have
> > problems (as hashcode and equals behaviors are defined as kind of "identity"
> > of the object).
> >
> > []s
> > Edson
> >
> > 2007/4/24, Chris West <crayzfishr at gmail.com>:
> > >
> > > Mark,
> > >
> > > The object being retracted currently has no hashcode and equals,
> > > therefore defaulting to the ones on Object. The reason I'm even retracting
> > > the object is so that I can assert it again to work around bug
> > > http://jira.jboss.com/jira/browse/JBRULES-614. Therefore the reason I
> > > am retracting the object is because it has changed.
> > >
> > > Is there a requirement to implement hashcode and equals? If so, does
> > > it have to be based on all fields of the pojo? My intent was to implement
> > > hashcode and equals, but only base them on the id fields of the class.
> > >
> > > -Chris
> > >
> > > On 4/24/07, Mark Proctor < mproctor at codehaus.org> wrote:
> > > >
> > > > Have you correctly implemented your equals and hashcode? Have you
> > > > made sure that the objects fields haven't changed between the assert and the
> > > > retract?
> > > >
> > > > Mark
> > > > Chris West wrote:
> > > >
> > > > Hi,
> > > >
> > > > Using version 3.0.6 of jboss rules, while retracting an object, I
> > > > get the following:
> > > >
> > > > java.lang.NullPointerException
> > > > at org.drools.reteoo.beta.InstanceEqualConstrRightMemory.remove(Unknown
> > > > Source)
> > > > at org.drools.reteoo.BetaMemory.remove (Unknown Source)
> > > > at org.drools.reteoo.NotNode.retractObject(Unknown Source)
> > > > at org.drools.reteoo.ObjectSource.propagateRetractObject(Unknown
> > > > Source)
> > > > at org.drools.reteoo.RightInputAdapterNode.retractTuple (Unknown
> > > > Source)
> > > > at org.drools.reteoo.TupleSource.propagateRetractTuple(Unknown
> > > > Source)
> > > > at org.drools.reteoo.NotNode.retractTuple(Unknown Source)
> > > > at org.drools.reteoo.TupleSource.propagateRetractTuple (Unknown
> > > > Source)
> > > > at org.drools.reteoo.JoinNode.retractObject(Unknown Source)
> > > > at org.drools.reteoo.ObjectSource.propagateRetractObject(Unknown
> > > > Source)
> > > > at org.drools.reteoo.ObjectTypeNode.retractObject (Unknown
> > > > Source)
> > > > at org.drools.reteoo.Rete.retractObject(Unknown Source)
> > > > at org.drools.reteoo.ReteooRuleBase.retractObject(Unknown
> > > > Source)
> > > > at org.drools.reteoo.ReteooWorkingMemory.doRetract(Unknown
> > > > Source)
> > > > at org.drools.common.AbstractWorkingMemory.retractObject(Unknown
> > > > Source)
> > > > at org.drools.common.AbstractWorkingMemory.retractObject(Unknown
> > > > Source)
> > > > at ascc.rules.AbstractRulesCoordinator.statusChanged(
> > > > AbstractRulesCoordinator.java :354)
> > > > at ascc.rules.AbstractRulesCoordinator$1.statusChanged(
> > > > AbstractRulesCoordinator.java:111)
> > > > at ascc.status.BaseStatusBoard.notifyChange(BaseStatusBoard.java
> > > > :211)
> > > > at ascc.status.BaseStatusBoard.notifyChange (
> > > > BaseStatusBoard.java:191)
> > > > at ascc.status.DefaultParkingStatusBoard.empty(
> > > > DefaultParkingStatusBoard.java:318)
> > > > at ascc.spotting.DefaultSpottingCoordinator.computeInitialSpot(
> > > > DefaultSpottingCoordinator.java:578)
> > > > at ascc.spotting.DefaultSpottingCoordinator.statusCreated(
> > > > DefaultSpottingCoordinator.java:827)
> > > > at ascc.status.BaseStatusBoard.notifyCreate(BaseStatusBoard.java
> > > > :100)
> > > > at ascc.status.BaseStatusBoard.notifyCreate (
> > > > BaseStatusBoard.java:80)
> > > > at
> > > > ascc.status.DefaultAirPlanStatusBoard.missionAssignmentComplete(
> > > > DefaultAirPlanStatusBoard.java:479)
> > > > at
> > > > ascc.missionassignment.Rule_6___Complete_mission_assignment_for_event_0.consequence(Rule_6___Complete_mission_assignment_for_event_0.java:36)
> > > > at
> > > > ascc.missionassignment.Rule_6___Complete_mission_assignment_for_event_0ConsequenceInvoker.evaluate(Rule_6___Complete_mission_assignment_for_event_0ConsequenceInvoker.java:24)
> > > >
> > > > ... 7 more
> > > >
> > > > Is there something I could be doing wrong? Is this a known bug?
> > > >
> > > > Thanks,
> > > > -Chris
> > > >
> > > > ------------------------------
> > > >
> > > > _______________________________________________
> > > > rules-users mailing list
> > > > rules-users at lists.jboss.org
> > > >
> > > >
> > > > https://lists.jboss.org/mailman/listinfo/rules-users
> > > >
> > > >
> > > >
> > > > _______________________________________________
> > > > rules-users mailing list
> > > > rules-users at lists.jboss.org
> > > > https://lists.jboss.org/mailman/listinfo/rules-users
> > > >
> > > >
> > >
> > > _______________________________________________
> > > rules-users mailing list
> > > rules-users at lists.jboss.org
> > > https://lists.jboss.org/mailman/listinfo/rules-users
> > >
> > >
> >
> >
> > --
> > Edson Tirelli
> > Software Engineer - JBoss Rules Core Developer
> > Office: +55 11 3529-6000
> > Mobile: +55 11 9287-5646
> > JBoss, a division of Red Hat @ www.jboss.com
> > _______________________________________________
> > rules-users mailing list
> > rules-users at lists.jboss.org
> > https://lists.jboss.org/mailman/listinfo/rules-users
> >
> >
>
> _______________________________________________
> rules-users mailing list
> rules-users at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
>
>
--
Edson Tirelli
Software Engineer - JBoss Rules Core Developer
Office: +55 11 3529-6000
Mobile: +55 11 9287-5646
JBoss, a division of Red Hat @ www.jboss.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/rules-users/attachments/20070425/9cd34ad3/attachment.html
More information about the rules-users
mailing list