Bug 1106300

Thiago,

I am investigating the issue you reported using your reproducer (thanks a lot for it). Indeed we have a bug there and I'll try to fix it asap.

In the meanwhile if you need a workaround for it you can try to replace the

eval ($cnt > 0)

after the accumulates with a 

exists Reading( type == $type ) 

before it.

Hello Mario, thanks for investigating it.

I tried the workaround you explained (see one of my rules below), but I had the same results.

Is there something else I should be doing?


rule "Abnormal"
dialect "mvel"

when
	threshold : Threshold ($type : type, $limHL : limitHardLow, $limSL : limitSoftLow, $limHH : limitHardHigh, $limSH : limitSoftHigh)
	exists Reading( type == $type ) from entry-point "entrypoint"
	accumulate( reading : Reading( type == $type ) over window:length( 10 ) from entry-point "entrypoint", $avg : average( reading.value ), $cnt : count( 1 ) )
	eval (($avg < $limSL && $avg >= $limHL) || ($avg > $limSH && $avg <= $limHH))
	not Alarm (type == $type, endDate == null)

then
	Alarm alarm = new Alarm();
	alarm.setType($type);
	alarm.setBeginDate(new java.util.Date());
        alarm.setLevel( "ABNORMAL" );
	insert(alarm);
	System.out.println("CREATED ABNORMAL ALARM - VALUE = " + $avg);
end

Sorry Thiago, you're right. I am still investigating the problem and I believe the bug is somewhat related with evals. So, in addition to this, you should also eliminate the 2nd eval. For instance you could replace it with a pattern matching like it follows.

rule "Abnormal"
dialect "mvel"

when
	threshold : Threshold ($type : type, $limHL : limitHardLow, $limSL : limitSoftLow, $limHH : limitHardHigh, $limSH : limitSoftHigh)
	exists Reading( type == $type ) from entry-point "entrypoint"
	Number( floatValue < $limSL && floatValue >= $limHL ) from accumulate( reading : Reading( type == $type ) over window:length( 10 ) from entry-point "entrypoint", average( reading.value ) )
	not Alarm (type == $type, endDate == null)

then
	Alarm alarm = new Alarm();
	alarm.setType($type);
	alarm.setBeginDate(new java.util.Date());
        alarm.setLevel( "ABNORMAL" );
	insert(alarm);
	System.out.println("CREATED ABNORMAL ALARM - VALUE = " + $avg);
end

Let me know if this works for you.

Mario

Mario, I removed the "evals" from my rules, and changed them all to use the pattern matching as you instructed, but now I am receiving the following error after the second fact insertion:

Exception in thread "Thread-0" java.lang.NullPointerException
	at org.drools.core.common.DoubleBetaConstraints.isAllowedCachedLeft(DoubleBetaConstraints.java:108)
	at org.drools.core.phreak.PhreakAccumulateNode.evaluateResultConstraints(PhreakAccumulateNode.java:710)
	at org.drools.core.phreak.PhreakAccumulateNode.doNode(PhreakAccumulateNode.java:89)
	at org.drools.core.phreak.RuleNetworkEvaluator.switchOnDoBetaNode(RuleNetworkEvaluator.java:562)
	at org.drools.core.phreak.RuleNetworkEvaluator.evalBetaNode(RuleNetworkEvaluator.java:533)
	at org.drools.core.phreak.RuleNetworkEvaluator.innerEval(RuleNetworkEvaluator.java:334)
	at org.drools.core.phreak.RuleNetworkEvaluator.outerEval(RuleNetworkEvaluator.java:161)
	at org.drools.core.phreak.RuleNetworkEvaluator.evaluateNetwork(RuleNetworkEvaluator.java:116)
	at org.drools.core.phreak.RuleExecutor.reEvaluateNetwork(RuleExecutor.java:194)
	at org.drools.core.phreak.RuleExecutor.evaluateNetworkAndFire(RuleExecutor.java:67)
	at org.drools.core.common.DefaultAgenda.fireNextItem(DefaultAgenda.java:935)
	at org.drools.core.common.DefaultAgenda.fireUntilHalt(DefaultAgenda.java:1160)
	at org.drools.core.common.AbstractWorkingMemory.fireUntilHalt(AbstractWorkingMemory.java:1002)
	at org.drools.core.common.AbstractWorkingMemory.fireUntilHalt(AbstractWorkingMemory.java:978)
	at org.drools.core.impl.StatefulKnowledgeSessionImpl.fireUntilHalt(StatefulKnowledgeSessionImpl.java:273)
	at br.com.redhat.testcase.Runner$1.run(TestCase.java:30)

Thiago, this seems to be a different issue and I cannot reproduce it. Can you please also attach your modified test case reproducing this problem?

Thanks,
Mario

Fixed by https://github.com/droolsjbpm/drools/commit/320f34fd0

Cherry-picked to 6.1.x branch with https://github.com/droolsjbpm/drools/commit/f6af4dbc1

Created attachment 907611 [details]
Project with a new test case

Mario, I am attaching "test-case2" project that reproduces this different issue (NPE at  org.drools.core.common.DoubleBetaConstraints.isAllowedCachedLeft). 

If this is the case, I can open a new BZ, but please let me know.

Thanks,

Thiago

Cherry-picked to 6.0.x branch with https://github.com/droolsjbpm/drools/commit/0f990e02b

Verified on BRMS 6.0.2 CR1.