[jboss-jira] [JBoss JIRA] Commented: (JBRULES-2881) sliding window:length(1) sticks to old event

[Edson Tirelli (JIRA)]

    [ https://issues.jboss.org/browse/JBRULES-2881?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12603453#comment-12603453 ] 

Edson Tirelli commented on JBRULES-2881:
----------------------------------------

Great catch Wolfgang. The problem was still there, as the engine was not correctly processing the behaviors for the first pattern in the rule. Now the result is the correct one:

BA: ( 2, 3 )
AB: ( 3, 2 )
BA: ( 1, 3 )
AB: ( 3, 1 )
BA: ( 1, 2 )
AB: ( 2, 1 )


> sliding window:length(1) sticks to old event
> --------------------------------------------
>
>                 Key: JBRULES-2881
>                 URL: https://issues.jboss.org/browse/JBRULES-2881
>             Project: Drools
>          Issue Type: Bug
>      Security Level: Public(Everyone can see) 
>          Components: drools-core, drools-core (fusion)
>    Affects Versions: 5.1.1.FINAL
>            Reporter: Wolfgang Laun
>            Assignee: Edson Tirelli
>            Priority: Critical
>             Fix For: 5.2.0.CR1
>
>
> A rule combining two events of type A and B, each "seen" through a sliding window of length 1 misfires. Rather that combining the latest A with the latest B, it keeps firing the latest B together with *all* old A's.
> Interestingly, this effect disappears if the order of patterns in the rule is inverted. Given the event sequence: A, A, B, A, B, the
>    rule ab when A() over window:length(1) B() over window:length(1) then ... end
> misfires, but 
>    rule ba when B() over window:length(1) A() over window:length(1) then ... end
> works correctly.
> This may be related to JBRULES-2862.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira