[ https://issues.jboss.org/browse/DROOLS-200?page=com.atlassian.jira.plugin... ] Pieter van den Brink commented on DROOLS-200: --------------------------------------------- I found another dataset for which the issue is still reproducible after correcting the hashcode and equals of FamilyStart (both in 6.0.0.Beta5 and 6.0.0-SNAPSHOT). See drools-200-isolated-reproducer-v2.zip, with the corrected hashcode/equals and a modified dataset created in ScheduleSolver's buildSchedule. When running the solver, the constraintMatchTotalMap of the ScoreDirector contains this IntConstraintMatchTotal at the time of corruption: [defaultpkg, familyCompact, 1]=defaultpkg/familyCompact/level1=-6 This IntConstraintMatchTotal's constraintMatchSet contains the following: defaultpkg/familyCompact/level1/[FamilyEnd Family 1, Timeslot 2, FamilyStart Family 1, Timeslot 1]=-1=java.lang.Object@4a58115f defaultpkg/familyCompact/level1/[FamilyEnd Family 1, Timeslot 3, FamilyStart Family 1, Timeslot 1]=-2=java.lang.Object@4a58115f defaultpkg/familyCompact/level1/[FamilyEnd Family 0, Timeslot 3, FamilyStart Family 0, Timeslot 0]=-3=java.lang.Object@4a58115f So it appears that there are two FamilyEnd's in memory for Family 1, which should not happen. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira