[rules-users] PatternBuilder createRestriction bug on Drools 5.1.1
Wolfgang Laun
wolfgang.laun at gmail.com
Thu Dec 22 05:29:41 EST 2011
Sorry - I should have made it clear that it doesn't matter *why* one
of the components of a composite restriction is in error. Either a
missing declaration or a type incompatibility - it results in a null
reference, and some agent processing this doesn't check.
-W
2011/12/22 Manuel Ortiz <manuel.ortizramos at gmail.com>:
> Dear Wolfgang:
>
> Sorry, I don't understand your answer very well. The bug I report has to do
> with composite restrictions, not with static variable declarations. Maybe my
> bug description is not very good.
>
> I used "Hello world" example described at
>
> http://downloads.jboss.com/drools/docs/5.1.1.34858.FINAL/drools-expert/html/ch08.html#d0e6981
>
> changing the first single restriction to a composite restriction in which
> the second part of the restriction compares an int to a string literal to
> fire a wrong type rule compile error. But as this second restriction belongs
> to a composite restriction, the composite one is created with a null
> component, no compile error reported.
>
> I understand that Message.HELLO is correctly declared in the "Hello World"
> example project and that this static variable shouldn't be a problem at
> compiling. In fact, if you change the static variable by a literal number,
> the error would be the same.
>
> Sorry for my english. I hope this explanation helps to understand the bug.
>
> Kind regards,
>
> Manuel Ortiz.
>
> 2011/12/22 Wolfgang Laun <wolfgang.laun at gmail.com>
>>
>> Using an undeclared static variable (Message.HELLO) wasn't handled
>> cleanly in 5.1.1. This has been fixed in 5.2.0.
>> -W
>>
>>
>> 2011/12/22 Manuel Ortiz <manuel.ortizramos at gmail.com>:
>> > Dear Wolfgang:
>> >
>> > Just change the "Hello World" rule in the "Hello World" example by this
>> > one...
>> >
>> > rule "Hello World"
>> > when
>> > m : Message( status == Message.HELLO || != "WrongType", myMessage :
>> > message
>> > )
>> > then
>> > System.out.println( myMessage );
>> > m.setMessage( "Goodbye cruel world" );
>> > m.setStatus( Message.GOODBYE );
>> > update( m );
>> > end
>> >
>> > ...and you will get the following building error:
>> >
>> > Description Resource Path Location Type
>> > Error: java.lang.NullPointerException Sample.drl
>> > /DroolsTest2/src/main/rules
>> > Unknown Drools Error
>> >
>> > This NullPointerException breaks the normal drools compiler execution
>> > flow.
>> > In contrast, if you use the following "Hello World" rule...
>> >
>> > rule "Hello World"
>> > when
>> > m : Message( status != "WrongType", myMessage : message )
>> > then
>> > System.out.println( myMessage );
>> > m.setMessage( "Goodbye cruel world" );
>> > m.setStatus( Message.GOODBYE );
>> > update( m );
>> > end
>> >
>> > ...then you get the following building errors:
>> >
>> > Description Resource Path Location Type
>> > BuildError: Unable to create a Field value of type 'ValueType = 'int''
>> > and
>> > value 'WrongType' Sample.drl /DroolsTest2/src/main/rules Unknown Drools
>> > Error
>> > BuildError: Unable to create restriction '[LiteralRestriction: !=
>> > WrongType]' for field 'status' in the rule 'Hello World' Sample.drl
>> > /DroolsTest2/src/main/rules line 7 Drools Error
>> >
>> > which don't break the drools compiler execution flow.
>> >
>> > I think this one should be the right output for the first example.
>> >
>> > Below you can find my fix to this problem, based on the assumption that
>> > createRestriction must return null if the restriction (neither composite
>> > nor
>> > simple) cannot be constructed.
>> >
>> > At PatternBuilder.java, substitute createRestriction by...
>> >
>> > private Restriction createRestriction(final RuleBuildContext
>> > context,
>> > final Pattern pattern,
>> > final FieldConstraintDescr
>> > fieldConstraintDescr,
>> > final
>> > RestrictionConnectiveDescr
>> > top,
>> > final InternalReadAccessor
>> > extractor) {
>> > Restriction[] restrictions = new
>> > Restriction[top.getRestrictions().size()];
>> > int index = 0;
>> >
>> > for ( Iterator it = top.getRestrictions().iterator();
>> > it.hasNext();
>> > ) {
>> > RestrictionDescr restrictionDescr = (RestrictionDescr)
>> > it.next();
>> >
>> > if ( restrictionDescr instanceof RestrictionConnectiveDescr
>> > ) {
>> > restrictions[index] = this.createRestriction( context,
>> > pattern,
>> >
>> > fieldConstraintDescr,
>> >
>> > (RestrictionConnectiveDescr) restrictionDescr,
>> >
>> > extractor );
>> >
>> > } else {
>> > restrictions[index] = buildRestriction( context,
>> > pattern,
>> > extractor,
>> >
>> > fieldConstraintDescr,
>> > restrictionDescr
>> > );
>> > if ( restrictions[index] == null ) {
>> > context.getErrors().add( new DescrBuildError(
>> > context.getParentDescr(),
>> >
>> > fieldConstraintDescr,
>> > null,
>> >
>> > "Unable to
>> > create restriction '" + restrictionDescr.toString() + "' for field '" +
>> > fieldConstraintDescr.getFieldName() + "' in the rule '" +
>> > context.getRule().getName() + "'" ) );
>> > }
>> > }
>> >
>> > if(restrictions[index] == null)
>> > {
>> > index = 0;
>> > break;
>> > }
>> > else
>> > {
>> > index++;
>> > }
>> > }
>> >
>> > if(index == 0) {
>> > return null;
>> > } else if ( restrictions.length > 1 ) {
>> > AbstractCompositeRestriction composite = null;
>> > if ( top.getConnective() == RestrictionConnectiveDescr.AND )
>> > {
>> > composite = new AndCompositeRestriction( restrictions );
>> > } else if ( top.getConnective() ==
>> > RestrictionConnectiveDescr.OR
>> > ) {
>> > composite = new OrCompositeRestriction( restrictions );
>> > } else {
>> > context.getErrors().add( new DescrBuildError(
>> > context.getParentDescr(),
>> >
>> > fieldConstraintDescr,
>> > null,
>> > "This is a
>> > bug: Impossible to create a composite restriction for connective: " +
>> > top.getConnective() + "' for field '" +
>> > fieldConstraintDescr.getFieldName()
>> > + "' in the rule '"
>> > +
>> > context.getRule().getName() + "'" ) );
>> > }
>> >
>> > return composite;
>> > } else if ( restrictions.length == 1 ) {
>> > return restrictions[0];
>> > }
>> > context.getErrors().add( new DescrBuildError(
>> > context.getParentDescr(),
>> >
>> > fieldConstraintDescr,
>> > null,
>> > "This is a bug:
>> > trying
>> > to create a restriction for an empty restriction list for field '" +
>> > fieldConstraintDescr.getFieldName() + "' in the rule '" +
>> > context.getRule().getName() + "'" ) );
>> > return null;
>> > }
>> >
>> > Kind regards,
>> >
>> > Manuel Ortiz.
>> >
>> >
>> > 2011/12/22 Wolfgang Laun <wolfgang.laun at gmail.com>
>> >>
>> >> Now would you mind posting an example reproducing this error...
>> >> -W
>> >>
>> >> On 22/12/2011, Manuel Ortiz <manuel.ortizramos at gmail.com> wrote:
>> >> > Dear Sirs:
>> >> >
>> >> > I'm working on an application which writes a set of rules using
>> >> > information
>> >> > which is stored in BDD tables. This information can be 'wrong'
>> >> > concerning
>> >> > rule compilation, and I expect that Drools reports the corresponding
>> >> > compilation errors.
>> >> >
>> >> > I've found that PatternBuilder.createRestriction() fails to process
>> >> > composite restrictions when one of the single restrictions cannot be
>> >> > implemented. The problem is that although one of the sigle
>> >> > restrictions
>> >> > is
>> >> > returned null, the composite restriction array is succesfully
>> >> > returned,
>> >> > one
>> >> > of its elements being null, wich causes a NullPointerException when
>> >> > the
>> >> > composite restriction is processed in
>> >> > PatternBuilder.build(AbstractCompositeConstraint version) to obtain a
>> >> > MultiRestrictionFieldConstraint object.
>> >> >
>> >> > I would like to know if this bug has yet been reported and fixed in
>> >> > some
>> >> > Drools version beyond 5.1.1
>> >> >
>> >> > Thank you very much for your time.
>> >> >
>> >> > Kind regards,
>> >> >
>> >> > Manuel Ortiz
>> >> >
>> >> _______________________________________________
>> >> 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
>> >
>>
>> _______________________________________________
>> 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
>
More information about the rules-users
mailing list