[rules-users] Constaint evaluation

[Greg Barton]

The ideas is that objects in working memory are those that meant to be tracked: their changes are made visible to the rules via the insert/update/retract methods.  A global is not in working memory, so it's changes cannot be tracked.  You can change the contents of a global all day long and the rules would never be notified.

It's a similar concept to threads and synchronization.  You can have unsynchronized access to member variables in a class when in a multithreaded environment, but the results are unpredictable.  Likewise, you can use globals in conditions, and change the value of the global as you go, but the results are unpredictable.  

--- On Wed, 6/10/09, Malenfant, Andre <andre.malenfant at cgi.com> wrote:

> From: Malenfant, Andre <andre.malenfant at cgi.com>
> Subject: RE: [rules-users] Constaint evaluation
> To: "Kris Verlaenen" <Kris.Verlaenen at cs.kuleuven.be>
> Cc: "Rules Users List" <rules-users at lists.jboss.org>
> Date: Wednesday, June 10, 2009, 2:20 PM
> Thanks Kris,
> 
> I will try with code constraints.
> 
> But on the globals subject, I read in the documentation
> that one should not use a global in a rule. I have trouble
> understanding why. I don't think drools clones the global
> object (might not be possible anyway). If my rule
> consequences change the state of a global (this usage is
> considered valid as per the documentation as well) then what
> is the harm of using it in a condition? I could understand
> with immutable objects like String but...
> 
> I can change my code to insert that global in the working
> memory but I would like to understand why I should do so...
> And if I do and the state of that object is changed by a
> consequence of my rules, I have to call update? For the same
> reason? I guess I am trying to understand the underlying
> mechanism.
> 
> Thanks
> 
> 
> -----Original Message-----
> From: Kris Verlaenen [mailto:Kris.Verlaenen at cs.kuleuven.be]
> 
> Sent: Wednesday, June 10, 2009 11:38 AM
> To: Rules Users List; Malenfant, Andre
> Cc: Rules Users List
> Subject: Re: [rules-users] Constaint evaluation
> 
> It depends on what type of constraint you are using.
> 
> If you are using a rule constraint, that constraint will be
> evaluated
> just the same as normal rules.  This means that the
> constraints are
> evaluated when data is inserted / updated / removed. 
> Note that, if you
> want to make sure the engine is using up-to-date
> information, you must
> notify the engine (using update) when you change the data
> in the working
> memory.  Also note that globals are considered
> immutable.  You should
> never write rules that depend on the state of a global and
> where the
> global can be changed.
> 
> If you are using code constraints, the constraint is
> evaluated at the
> point the code constraint is reached.
> 
> Kris
> 
> Quoting "Malenfant, Andre" <andre.malenfant at cgi.com>:
> 
> > I have some trouble understanding how conditions and
> constraints are
> > evaluated:
> > 
> > I have a rule flow split node with constraints on a
> global object.
> > The split node always takes the same path like if the
> constraints are
> > evaluated at the beginning of the execution of the
> process and not
> > when the process flow reaches that split node. What I
> would expect is
> > that the split node constraints takes into account the
> state of the
> > global object as modified by previous rules in the
> flow.
> > 
> > Am I right to think that rules conditions and split
> nodes constraints
> > are evaluated only when inserting/updating objects in
> the working
> > memory? The documentation is not really useful (unless
> I haven't
> > found the appropriate one).
> > 
> > Thanks
> > 
> > André
> > 
> > -----Original Message-----
> > From: rules-users-bounces at lists.jboss.org
> > [mailto:rules-users-bounces at lists.jboss.org]
> On Behalf Of Malenfant,
> > Andre
> > Sent: Wednesday, June 10, 2009 11:07 AM
> > To: Kris Verlaenen
> > Cc: Rules Users List
> > Subject: RE: [rules-users] Globals in ruleflow
> > 
> > Thanks for trying this for me...
> > 
> > As it turns out, while I was creating a test sample
> for you I
> > realized
> > that my ruleflow was not loading properly and I forgot
> to check for
> > errors on the builder. Now it works.
> > 
> > Still, the behavior is strange. If my rule doesn't
> load and even
> > though
> > my DRL loaded, my globals were not available. Since
> the DRL declared
> > the
> > same globals it should not have given me that error.
> > 
> > Thanks
> > 
> > -----Original Message-----
> > From: Kris Verlaenen [mailto:Kris.Verlaenen at cs.kuleuven.be]
> 
> > Sent: Tuesday, June 09, 2009 5:51 PM
> > To: Rules Users List; Malenfant, Andre
> > Cc: Rules Users List
> > Subject: Re: [rules-users] Globals in ruleflow
> > 
> > Andre,
> > 
> > There should be no problem in using the same global in
> both your
> > rules
> > and processes.
> > 
> > I have tried a simple example as you described but
> have not been able
> > to
> > reproduce the problem.  Could you send me a
> self-contained example
> > that
> > shows the issue?
> > 
> > Kris
> > 
> > Quoting "Malenfant, Andre" <andre.malenfant at cgi.com>:
> > 
> > > I am experimenting with rule flows and I get the
> following error:
> > > 
> > > Unexpected global [myglobal]
> > > 
> > > When calling setGlobal on the session.
> > > 
> > > This code works without the workflow (globals
> declared in the drl)
> > > but
> > > fails when I include the rule flow. I declared
> the same globals in
> > > the
> > > rule flow in the header section.
> > > 
> > > _______________________________________________
> > > rules-users mailing list
> > > rules-users at lists.jboss.org
> > > https://lists.jboss.org/mailman/listinfo/rules-users
> > > 
> > 
> > 
> > 
> > 
> > Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm
> > 
> > _______________________________________________
> > 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
> > 
> 
> 
> 
> 
> Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm
> 
> _______________________________________________
> rules-users mailing list
> rules-users at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
>