[hibernate-dev] Is SQM going to remain a separate component?

[Sanne Grinovero]

Thanks Gunnar, I'm wondering the same.

At the same time though, I would rather avoid the work of adapting the
SQM code to an ORM version which was not meant to be compatible with
it.
I'd rather wait with that until we base OGM on ORM6 and invest the
saved energy otherwise.

Mine is not a technical suggestion, but just an observation that such
work would only benefit very selected OGM releases: it doesn't bring
long term value.
If we have spare cycles to dedicate on the parser work (which I doubt
anyway) I'd rather us start working on the ORM6 previews as soon as
there's a testable branch to make sure that future integration will be
great.

Thanks,
Sanne


On 27 May 2016 at 09:01, Gunnar Morling <gunnar at hibernate.org> wrote:
> Steve, all,
>
> During the No-ORM meeting this week, Andrea gave us a great overview on the
> current status of the SQM work (thanks again, Andrea!).
>
> This looks awesome and we look forward to make use of it in Hibernate OGM.
> One thing I'm wondering is how closely tied to Hibernate ORM 6 this will
> be. Is the plan to move all the SQM bits into the hibernate-orm repo
> eventually? Or will SQM remain a separate component, and only the
> implementation of the type system will live in ORM?
>
> The reason for asking is that I'm trying to gauge how feasible it'd be to
> use SQM in a Hibernate OGM based on ORM 5.x. That'd be much simpler (or
> even only actually doable without essentially copying code), if SQM would
> remain separate, and we'd "only" have to copy/backport the type system
> implementation.
>
> Any thoughts?
>
> Thanks,
>
> --Gunnar
> _______________________________________________
> hibernate-dev mailing list
> hibernate-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/hibernate-dev