[jboss-jira] [JBoss JIRA] (JBRULES-3662) Rule activation error using length sliding window

[Davide Sottara (JIRA)]

    [ https://issues.jboss.org/browse/JBRULES-3662?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12770508#comment-12770508 ] 

Davide Sottara commented on JBRULES-3662:
-----------------------------------------

After trying to fix 3201 in 5.5.1-SNAPSHOT, I can run the NewTest without errors.
                
> Rule activation error using length sliding window
> -------------------------------------------------
>
>                 Key: JBRULES-3662
>                 URL: https://issues.jboss.org/browse/JBRULES-3662
>             Project: JBRULES
>          Issue Type: Bug
>      Security Level: Public(Everyone can see) 
>          Components: drools-core
>    Affects Versions: 5.3.0.Final, 5.4.0.Final, 5.5.0.Beta1
>         Environment: Mac OS X, running Apple Java 1.6, Maven2, Eclipse Juno 
>            Reporter: Vítor Moreira
>            Assignee: Mark Proctor
>             Fix For: 6.0.0.Alpha1
>
>         Attachments: droolsTest.zip
>
>
> When using "window:length", some rules are activated twice, when it should be only activated once.
> The uploaded example reproduces the error: we insert Transactions, which has a field "problem" inside: 1 for problem, 0 for OK. The rules calculates the percent of transactions with problems. If the percentage>=40%, raises an alarm. If it's below 40%, removes the alarm (if there's an alarm). There is a minimum of 5 transactions to trigger/rearm.
> The test case simulates a raise to 40% (alarm is trigged), continues high (until 60%) and starts to fall down to 20% (alarm is removed).
> Drools version 5.2.0.Final: alarm is trigged *once*, removed *once*
> Drools version 5.3.0.Final: alarm is trigged *once*, removed *twice*
> Drools version 5.4.0.Final: alarm is trigged *once*, removed *twice*
> Drools version 5.5.0.Beta1: alarm is trigged *once*, removed *twice*
> Since there's only one alarm and drools tries to remove twice, there's a NullPointerException.
> *NOTE:* the attached zip file has code to test issue 3564 because I first though that could be a reopen issue but no, it's not the same issue (issue 3564 is resolved in 5.5.0.Beta1).
> *NOTE2:* I don't know if it's related to issue 3201...

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira