I was able to use the Drools Eclipse Plugin to view the Rete tree and I can
confirm that it Drools does build the tree differently under the scenarios
described above. With rule 3 commented out, there are 2 AlphaNodes -- one
checks for [role=="1"] and the other checks for [role=='1']. With rule
3
uncommented, the AlphaNode for [role=='1'] disappears. It seems rather
scary that Drools doesn't create the Rete tree the same way. I hope there
aren't any other issues like this that may affect the validity of the
rules???
Do you recommend skipping the upgrade to Drools 5.4.0 in favor of another
version?
--
View this message in context:
http://drools.46999.n3.nabble.com/Drools-Rule-Evaluations-are-Inconsisten...
Sent from the Drools: User forum mailing list archive at
Nabble.com.