[
https://issues.apache.org/jira/browse/OPENJPA-2519?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14073224#comment-14073224
]
Rick Curtis commented on OPENJPA-2519:
--------------------------------------
Marc -
Can you describe your scenario a bit better? It sounds like your scenario
roughly is :
* Find/load an Entity(e1). Somehow it gets detached (and then serialized?).
This Entity has a Lazy OneToOne
* Set the OneToOne relationship to a new/existing instance(e2).
* Call em.merge(e1) and expect that the relationship for e1->e2 is persisted?
It would also be helpful if you posted the contents of your persistence.xml.
Hopefully with a little more information I can explain nearly everything that
you're seeing.
I'm still at a loss why MySQL vs MariaDB would behave differently in this case.
> fetchType.lazy and persisting one-to-one relations
> --------------------------------------------------
>
> Key: OPENJPA-2519
> URL: https://issues.apache.org/jira/browse/OPENJPA-2519
> Project: OpenJPA
> Issue Type: Bug
> Components: jpa
> Affects Versions: 2.2.0
> Reporter: Marc Logemann
>
> Hi,
> i have an entity which has a mapped superclass. In that superclass i have a
> 1:1 relation which is marked "lazy".
> Now when i load that entity, attach an object to that relation and persist
> it, the change is not reflected in the database. But the jpaversion is
> increased. Also all other "normal" mapped fields are persisted but not the
> Lazy 1:1. Of course i checked the object right before persist and it looks
> pretty perfect.
> If i remove the lazy attribute from the annotation. Everything works like
> expected. When looking at OPENJPA-2505 and this issue, i am really concerned
> about your FetchType implementation. This definitely feels buggy.
> And to make things more weird: On my MySQL box, the bug is non-existant. As
> soon as i deploy the application (really the very same application) to
> MariaDB, its there.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
