[hibernate-dev] Hibernate ORM 4.1.1 release - Undocumented behavior change in Optimistic Versioning ?

[Steve Ebersole]

The old behavior violated JPA spec.  The issue is HHH-7138


On Fri 09 Mar 2012 03:04:54 AM CST, Demetz, Guenther wrote:
> Hi hibernate developers,
>
> testing Hibernate ORM 4.1.1 I discovered a behavior change regarding Optimistic Lock Versioning
> in detail on bidirectional OneToMany relations with inverse-owner (mappedBy attribute specified):
>
> - until Hiberante version 4.1.0 by default on changes both involved sides got a version increment
> - now with Hiberante4.1.1 only the owner-side (=the Many-side) gets a version increment.
>
> To make the thing working as before,
> now on such relations we must define explicitly OptimisticLock.excluded=false
>
> @javax.persistence.OneToMany(mappedBy = "parent")
> @org.hibernate.annotations.OptimisticLock(excluded=false)
> @javax.persistence.MapKey(name="field1")
> private Map<String,Child>  aggChild = new java.util.HashMap();
>
> Was such behavior change intended?
> I looked at the 28 issues which had been fixed with Hiberante4.1.1 but found nothing in regard,
> maybe I overlooked something ?
>
>
> Würth Phoenix S.r.l.
> Günther Demetz
> Via Kravogl 4
> I-39100 Bolzano
> E-Mail:     guenther.demetz at wuerth-phoenix.com
> Website:  www.wuerth-phoenix.com
>
>

-- 
steve at hibernate.org
http://hibernate.org