Hello there. I seem to have trouble with the "timer" operator which is documented to have superseded the "duration" operator for use in the rule body. I might have something to do with the "not" operator I use in my rule. Scenario: A is a fact, B is an event When I write something like this rule "test" timer(int: 5s 5s) when $a : A() not (exists B(a == $a)) then insert(new B($a)); // which means in the new B object: B.a = $a end

I expect this rule to fire every 5 seconds if and only if the LHS is true. I just insert a few A objects at the beginning. So I expect this rule to fire ONCE, because it will insert the B event which makes this rule's LHS evaluate to false in the future. BUT, reality is, this rule fires every 5 seconds, no matter what. Using the deprecated duration operator with "duration(5s)" instead of the timer operator, it works as expected. I assume, that timer works differently compared to duration. So I conclude that duration should coexist. Or am I missing something here on the usage of the timer operator? Thanks in advance. Oliver

Thanks, unfortunately not working, either. Maybe it will help, if I describe the scenario more detailed. I have three A facts inserted at the beginning. Then the rule fires the first time after 5 seconds for every A fact and inserts the B events. Then I retract ONE of the B events, but the rule keeps firing for EVERY A fact, despite the B events remaining in the WorkingMemory. -- View this message in context: http://drools-java-rules-engine.46999.n3.nabble.com/timer-and-duration-tp... Sent from the Drools - User mailing list archive at Nabble.com. _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users

On 09/12/2010 14:50, Wolfgang Laun wrote: I can confirm that the timer rule continues to fire - if the condition contains a not CE - even after the not CE should return false Repeating timer firing stops for "positive" CEs turning to false. sounds like a bug :( Wolfgang, could you commit your test to TimerAndCalendarTest. Take a look at the tests there and follow a similar format, generally when there is only one or two small rules we prefer to inline the text rather than externalise it in a file, see testIntervalTimer So that it doesn't break hudson add FIXME to the start of the rule name, such as FIXME_testTimerWithNot Ideally open a jira and when you commit the code put the jira id in the commit messgae along with the jira title and a message something like "adding failing unit test" Thanks Mark -W On 9 December 2010 15:06, OlliSee <o.roess(a)seeburger.de&gt; wrote: > > Thanks, > unfortunately not working, either. > > Maybe it will help, if I describe the scenario more detailed. > > I have three A facts inserted at the beginning. > Then the rule fires the first time after 5 seconds for every A fact and > inserts the B events. > Then I retract ONE of the B events, but the rule keeps firing for EVERY A > fact, despite the B events remaining in the WorkingMemory. > -- > View this message in context: > http://drools-java-rules-engine.46999.n3.nabble.com/timer-and-duration-tp... > Sent from the Drools - User mailing list archive at Nabble.com. > _______________________________________________ > rules-users mailing list > rules-users(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/rules-users _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users

On 10/12/2010 08:44, Wolfgang Laun wrote: > Done. There are two FIXME-tests at the end, one for timer() + not; the other one > demonstrating that drools.halt() from a rule with a timer() doesn't > terminate the > session - it also keeps firing. Notice that even with no timer() on > the "halt" rule > but with another rule having a periodic timer "overrules" the drools.halt(). Did you commit? as I'm not seeing anything in trunk. http://fisheye.jboss.org/changelog/JBossRules/trunk/drools-compiler/src/t...

Mark > -W > > > 2010/12/10 Mark Proctor&lt;mproctor(a)codehaus.org&gt;: >> On 09/12/2010 14:50, Wolfgang Laun wrote: >> >> I can confirm that the timer rule continues to fire >> - if the condition contains a not CE >> - even after the not CE should return false >> >> Repeating timer firing stops for "positive" CEs turning to false. >> >> sounds like a bug :( >> >> Wolfgang, could you commit your test to TimerAndCalendarTest. Take a look at >> the tests there and follow a similar format, generally when there is only >> one or two small rules we prefer to inline the text rather than externalise >> it in a file, see testIntervalTimer >> >> So that it doesn't break hudson add FIXME to the start of the rule name, >> such as >> FIXME_testTimerWithNot >> >> Ideally open a jira and when you commit the code put the jira id in the >> commit messgae along with the jira title and a message something like >> "adding failing unit test" >> >> Thanks >> >> Mark >> >> -W >> >> >> On 9 December 2010 15:06, OlliSee&lt;o.roess(a)seeburger.de&gt; wrote: >>> Thanks, >>> unfortunately not working, either. >>> >>> Maybe it will help, if I describe the scenario more detailed. >>> >>> I have three A facts inserted at the beginning. >>> Then the rule fires the first time after 5 seconds for every A fact and >>> inserts the B events. >>> Then I retract ONE of the B events, but the rule keeps firing for EVERY A >>> fact, despite the B events remaining in the WorkingMemory. >>> -- >>> View this message in context: >>> http://drools-java-rules-engine.46999.n3.nabble.com/timer-and-duration-tp... >>> Sent from the Drools - User mailing list archive at Nabble.com. >>> _______________________________________________ >>> rules-users mailing list >>> rules-users(a)lists.jboss.org >>> https://lists.jboss.org/mailman/listinfo/rules-users >> _______________________________________________ >> rules-users mailing list >> rules-users(a)lists.jboss.org >> https://lists.jboss.org/mailman/listinfo/rules-users >> >> >> _______________________________________________ >> rules-users mailing list >> rules-users(a)lists.jboss.org >> https://lists.jboss.org/mailman/listinfo/rules-users >> >> > _______________________________________________ > rules-users mailing list > rules-users(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/rules-users > > _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users

Hello folks. Got a new characteristic unveiled,... if I have a rule like this (with Fusion in STREAM mode)... (Sensor is a fact) rule "Test" timer(int: 7s 7s) when $sensor : Sensor() Number($avg : doubleValue > 0.0) from accumulate (LoopDetectorEvent($vel : velocity, sensor == $sensor) over window:time(7s), average($vel)) then ... end ...Drools fires the rule correctly every 7 seconds for each Sensor fact. BUT, it sometimes fires twice for random Sensors. This is of course not intended. Is this a bug or can this behaviour be explained? -- View this message in context: http://drools-java-rules-engine.46999.n3.nabble.com/timer-and-duration-tp... Sent from the Drools - User mailing list archive at Nabble.com. _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users

No I don't modify any LoopDetectorEvent.Actually, in some situations the rule fires even more than two times for some Sensors. I don't use the logger. I just do System.out.println's to see when a rule fires. . If you need any more info, then I will use the logger for you. What information might be found there? -- View this message in context: http://drools-java-rules-engine.46999.n3.nabble.com/timer-and-duration-tp... Sent from the Drools - User mailing list archive at Nabble.com. _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users