[hibernate-issues] [Hibernate-JIRA] Commented: (HHH-2439) delay IDENTITY insertions in the case of FlushMode.MANUAL/NEVER

[Steve Ebersole (JIRA)]

    [ http://opensource.atlassian.com/projects/hibernate/browse/HHH-2439?page=comments#action_26233 ] 

Steve Ebersole commented on HHH-2439:
-------------------------------------

Certainly it is a subtle change that has potential to break existing applications.  

But are those existing applications relying on questionable behavior?   I'd argue yes.  Would you agree that what we do now is correct for out-of-transaction operation?  So why not extended to FlushMode.MANUAL?

As for the app relying on the id value being available, then those apps are clearly not relying on a spec defined behavior.  I'd argue that they were relying on a quirk in our implementation.  As you outline so well in the book, according to JPA spec, they should not rely on a generated id being available in memory until the next flush cycle which incorporates the entity.

> delay IDENTITY insertions in the case of FlushMode.MANUAL/NEVER
> ---------------------------------------------------------------
>
>          Key: HHH-2439
>          URL: http://opensource.atlassian.com/projects/hibernate/browse/HHH-2439
>      Project: Hibernate3
>         Type: Improvement

>   Components: core
>     Reporter: Steve Ebersole
>     Assignee: Steve Ebersole
>      Fix For: 3.2.3

>
>
> Much like we do now in 3.2 for out of transaction operations, we should delay performing insertions for post-insert generators with FlushMode.isManualFlushMode

-- 
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.jspa
-
For more information on JIRA, see:
   http://www.atlassian.com/software/jira