[ http://opensource.atlassian.com/projects/hibernate/browse/HHH-2350?page=c... ] Thomas Champagne commented on HHH-2350: --------------------------------------- Hi When do you think this issue will be fixed ? Thomas -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators.... - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://opensource.atlassian.com/projects/hibernate/browse/HHH-2350?page=c... ] Gail Badner updated HHH-2350: ----------------------------- Description: When using "non-inverse" one-to-many relations as described in "2.4.6.2.3. Bidirectional association with indexed collections" (http://www.hibernate.org/hib_docs/annotations/reference/en/html/entity.ht...) hibernate is wrong when activating 2nd level cache. See the attached example (unzip in /usr/local/hibernate-annotations-3.2.1.GA, testname=org.hibernate.test.annotations.invonetomany.InverseOneToManyTest). More details can be found here: http://forum.hibernate.org/viewtopic.php?t=969456 (and also in InverseOneToManyTest.java). was: When using "inverse" one-to-many relations as described in "2.4.6.2.3. Bidirectional association with indexed collections" (http://www.hibernate.org/hib_docs/annotations/reference/en/html/entity.ht...) hibernate is wrong when activating 2nd level cache. See the attached example (unzip in /usr/local/hibernate-annotations-3.2.1.GA, testname=org.hibernate.test.annotations.invonetomany.InverseOneToManyTest). More details can be found here: http://forum.hibernate.org/viewtopic.php?t=969456 (and also in InverseOneToManyTest.java). Component/s: (was: core) caching (L2) This problem really involves a non-inverse association. I've corrected this in the title and description. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators.... - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://opensource.atlassian.com/projects/hibernate/browse/HHH-2350?page=c... ] Gail Badner updated HHH-2350: ----------------------------- Fix Version/s: 3.6 3.5.x -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators.... - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://opensource.atlassian.com/projects/hibernate/browse/HHH-2350?page=c... ] Gail Badner commented on HHH-2350: ---------------------------------- Here is what is happening: 1) The many-to-one association is (correctly) mapped with update="false" 2) PropertyFactory.buildStandardProperty(...) indicates that the many-to-one association is not dirty-checkable, unless the many-to-one association is mapped with not-found="ignore". (It passes alwaysDirtyCheck || property.isUpdateable() to the StandardProperty constructor; alwaysDirtyCheck is false unless not-found="ignore"). 3) When the entity with an update to the many-to-one association is flushed, TypeFactory.findDirty(...) does not dirty-check the association because the following is false: final boolean dirty = currentState[i] != LazyPropertyInitializer.UNFETCHED_PROPERTY && properties[i].isDirtyCheckable( anyUninitializedProperties ) && properties[i].getType().isDirty( previousState[i], currentState[i], includeColumns[i], session ); 4) The many-to-one association is (correctly) updated in the DB when the dirty non-inverse collection is updated by CollectionUpdateAction, which evicts the collection from the second-level cache 5) EntityUpdateAction is responsible for updating/removing a modified entity from the cache. Since the entity is not dirtied by other updated properties, an EntityUpdateAction is not created, and so the unmodified entity remains in the cache. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators.... - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://opensource.atlassian.com/projects/hibernate/browse/HHH-2350?page=c... ] Gail Badner updated HHH-2350: ----------------------------- Fix Version/s: (was: 3.5.x) 3.5.4 -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators.... - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://opensource.atlassian.com/projects/hibernate/browse/HHH-2350?page=c... ] Gail Badner resolved HHH-2350. ------------------------------ Resolution: Fixed Fix Version/s: (was: 3.6.x) 3.6.0.Beta2 Fixed in Branch_3_5 and trunk. This was fixed by changing ManyToOneType.isAlwaysDirtyChecked() to return true. If an entity has only its "inverse" many-to-one association that is dirty, its version will be incremented. This was not being done before this fix. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators.... - For more information on JIRA, see: http://www.atlassian.com/software/jira