I'm at my wits end. I have two simple rules in two separate packages, they are of the form: rule1:  import somepackage.*  global somepackage.Foo foo  ...  then     foo.doSomething(...)  end rule2:  import otherpackage.*  global otherpackage.Gar gar  ...  then     gar.doSomething(...)  end When mvel is resolving foo for rule1 it does so via org.mvel2.compiler.CompiledExpression#getValue:line-104 When resolving gar for rule2 it does so via org.mvel2.compiler.CompiledExpression#getValue:line-107 The difference is that for rule1 the variable 'importInjectionRequired' is true, which results in 'foo' being resolved, but for rule2 this variable is false which results in 'gar' not being resolved. I suspecting that its important that the packages and the global instances are added dynamically *after* the session is created (in scala'ish dialect): val session = createSessionUsingNoPackages val kbuilder1 = KnowledgeBuilderFactory.newKnowledgeBuilder kbuilder1.add("rule1.drl", ResourceType.DRL) session.knowledgeBase addKnowledgePackages (kbuilder1.getKnowledgePackages) session addGlobal ("foo", foo) val kbuilder2 = KnowledgeBuilderFactory.newKnowledgeBuilder kbuilder2.add("rule2.drl", ResourceType.DRL) session.knowledgeBase addKnowledgePackages (kbuilder2.getKnowledgePackages) session addGlobal ("Gar", gar) If just one or the other packages are added the session, then either variable is resolved. But if both are added the session the 'gar' is not resolved. It does not matter what order they added to the session. I can also try to access 'gar' in rule1 and it also will not be resolved. I guess I really don't expect any help on this, but I will boil it down to a simple test and submit it to jira -barry -- View this message in context: http://n3.nabble.com/MVEL-wows-what-does-importInjectionRequired-mean-tp6... 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

Oh boy, what a ride. In my real code the two variables are called "DeviceOperatingStateModel" and "DeviceUtilizationStateModel". Just as an experiment I renamed the second to "aa" -- low and behold the variable was resolved. Then I changed to "DeviceUtilizationStateModelx" and "deviceUtilizationStateModel" and both of those were resolved. This is so strange. There something about the name "DeviceUtilizationStateModel" that mvel just does not like! (Ok, surely that's not it and its something else in classpath. But its more fun to make believe that mvel doesn't like my names).