On Fri, Jan 16, 2009 at 7:16 AM, Scott Ferguson <ferg(a)caucho.com> wrote:
On Jan 15, 2009, at 5:42 AM, Gavin King wrote:
> OK, after thinking this through over the last couple of days I've come
> to the conclusion that we can live with something slightly less
> general. The usecase for this stuff is really state synchronization
> and therefore N-of-N semantics is the only thing that makes sense, and
> supporting guaranteed delivery sounds to me like total overkill.
Is it possible to reuse the @Asynchronous annotation from EJB 3.1 with the
exact same semantics? They seem essentially the same to me, and having a
shared definition for that aspect would be fantastic (especially from a
documentation perspective.) Having two very-similar but different
annotations may be hard to keep track of. Plus the current EJB definition
is more complete.
I will ask Ken about that. Note that EJB also has
@BeforeCompletion/@AfterCompletion annotations which are also very
similar to our @BeforeTransactionCompletion/@AfterTransactionCompletion.
There are a couple of minor mismatches:
* the EJB annotations are in javax.ejb instead of javax.annotations
* they don't have @AfterTransactionSuccess/@AfterTransactionFailure
(instead a boolean value is passed, which doesn't really work for us
* ours are currently specified on the parameter instead of the method
But none of these differences seems like anything really killer.
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org