[hibernate-dev] HHH-13317

[Luis Barreiro]

Hello Gail,

The code in question deals with the detection of bi-directional 
associations and only affects that feature.

What happens currently is that in certain situations fields are matched 
(based on the value of mappedBy) but the types don't actually match, 
leading to either the failure of enhancement or the generated code is 
invalid.

Again, this has nothing to do with lazy loading or dirty tracking.

Regards,

Luis Barreiro

Middleware Performance Team

<https://red.ht/sig>

On 9/9/19 10:55 PM, Gail Badner wrote:
> Hi Luis,
>
> IIUC, HHH-13446 deals with a mappedBy="...." value that is invalid. 
> I'm confused why the test in the PR is passing. Shouldn't an exception 
> be thrown?
>
> HHH-13317 deals with a mappedBy value that is an association in an 
> embeddable, which is valid.
>
> I know that Hibernate does not support a lazy association in an 
> embeddable with bytecode enhancement. IIRC, we expect the association 
> in the embeddable will not be lazy. It shouldn't thrown an exception 
> though.
>
> Luis, could you please shed some light on this?
>
> Thanks,
> Gail
>
> On Thu, Sep 5, 2019 at 5:14 PM Gail Badner <gbadner at redhat.com 
> <mailto:gbadner at redhat.com>> wrote:
>
>     I'll take a look. Thanks Luis!
>
>
>
>     On Thu, Sep 5, 2019 at 3:02 PM Luis Barreiro <lbarreiro at redhat.com
>     <mailto:lbarreiro at redhat.com>> wrote:
>
>         Hello Gail,
>
>         Without looking with much detail, this seems to be the same as
>         HHH-13446.
>
>         There is a PR for that for some time now:
>         https://github.com/hibernate/hibernate-orm/pull/2919
>
>         If it's the same issue, it should be fixed. While Javassist
>         fails when enhancing the classes, ByteBuddy succeeds but the
>         resulting classes are invalid.
>
>         Regards,
>
>         Luis Barreiro
>
>         Middleware Performance Team
>
>         <https://red.ht/sig>
>
>         On 9/5/19 10:54 PM, Gail Badner wrote:
>>         HHH-13317 is specifically a bug with bytecode enhancement using Javassist.
>>
>>         It showed up when running a test added for HHH-13241.
>>
>>         I've verified that the same issue affected 5.1, so this is not a regression.
>>
>>         Is it worthwhile to fix this issue in 5.3/5.4? Will Javassist continue to
>>         be supported in Hibernate ORM 6? If not, I'd like to close HHH-13317 as
>>         "Won't fix".
>>
>>         Thanks,
>>         Gail
>>         _______________________________________________
>>         hibernate-dev mailing list
>>         hibernate-dev at lists.jboss.org  <mailto:hibernate-dev at lists.jboss.org>
>>         https://lists.jboss.org/mailman/listinfo/hibernate-dev
>