On 12/11/2015 11:14 AM, Lukas Jungmann wrote:
> Hi Scott,
>
> On 12/11/15 4:33 PM, Scott Marlow wrote:
>> Hi,
>>
>> Just to be more clear about why I am disappointed about not yet being
>> able to participate in discussions about what will be included in the
>> JPA 2.2 (MR). I expected there to be earlier discussion about which
>> items are being added to the initial JPA 2.2 draft specification. I
>> also expected the discussion about which items (JPA spec jiras) to
>> include, to occur openly on an expert group mailing list. Instead, I
>> happened to read on the EE mailing list, that a plan of possible items
>> that will be under consideration for JPA 2.2, are available via [4]
>> (as well as [2]).
>
> The list of items for 2.2 is still open. Do you have anything for 2.2
> you want to have included and wasn't explicitly mentioned anywhere?
> From what I got to know you're interested mainly in better and/or
> tighter JPA/CDI integration, which I'm seeing to be quite important, right?
Yes, I think that
https://java.net/jira/browse/JPA_SPEC-83 is about
better JPA/CDI integration (e.g. allow EntityListener class to inject
bean that introduces cycle back on the same PU associated with the
EntityListener). We have been experimenting with some Hibernate/WildFly
changes (using javax.enterprise.inject.spi.AfterDeploymentValidation) to
register the entity listeners a bit later but still early enough during
the deployment cycle, that we don't have to wait until (concurrent)
application requests are being made. I haven't responded yet here, as
we have more to discuss, as to using
javax.enterprise.inject.spi.AfterDeploymentValidation or something like
that. In the interest of showing the world, that there is interest in
solving problems like this, I decided to share our progress (briefly)
here.
Another one that we discussed before on the JPA 2.1 eg list, is
https://java.net/jira/browse/JPA_SPEC-70 which is about introducing a
2-phase approach to deploying the persistence unit. Others were also
interested in this change, as it could simplify some of the JPA
container deployer logic.