jsr345-experts@ejb-spec.java.net

[jsr345-experts] Re: (EJB_SPEC-13) EJB 3.2 Lite Support for Asynchronous

From: Marina Vatkina <marina.vatkina_at_oracle.com>
Date: Fri, 30 Sep 2011 13:09:50 -0700

Antonio Goncalves wrote:
>
>
> On Wed, Sep 28, 2011 at 20:42, Marina Vatkina
> <marina.vatkina_at_oracle.com <mailto:marina.vatkina_at_oracle.com>> wrote:
>
> We have also @Suspend in JAX-RS that can't use Future return type
> as it'll block the client.
>
> Timeout:
> 1) I do not see asyncTimeout in @WebServlet - was it proposed and
> dropped?
>
>
> Yes, you are right. I just checked servlet 3.0 final and this is
> what's written :
>
> /A.1 Changes since Servlet 3.0 Proposed Final/
> /1. Re-factored some of the Async APIs - moved addAsyncListener to
> AsyncContext and renamed it to addListener. Moved setAsyncTimeout to
> AsyncContext and renamed it to setTimeout./

Thanks.

This is actually another point that I wanted to ask: why wouldn't EJB
async have an AsyncListener?
>
>
>
> 2) What time period will it define?
> a) to cancel if not started, or
> b) cancel after running for that long, or
> c) cancel of a combination of a) and b) is too long?
> Anything other than a) might not be even possible in XA transactions.
>
>
> You bring another point by mentioning XA transactions. If we take the
> transaction service out of the EJBs and make it available to managed
> beans, would that mean that managed bean will then be able to do XA
> transactions ?

Yes.
-marina
>
>
>
> -marina
>
> Antonio Goncalves wrote:
>
> No, asynch servlet does not really implement the comet style,
> it's just an easy way to do asynchronous calls in a different
> thread. Like Pete I think it's a shame to have different ways
> of doing asynchrnoy (asyncSupported = true in Servlet and
> Asynchronous elsewhere)... but I'm sure there's a reason. What
> asynch servlets also have is a timeout (asyncSupported = true,
> asyncTimeout = 3000). It could also be interesting to have a
> timeout attribute on @Asynchronous, what do you think ?
>
> On Tue, Sep 27, 2011 at 22:50, Rick Hightower
> <richardhightower_at_gmail.com
> <mailto:richardhightower_at_gmail.com>
> <mailto:richardhightower_at_gmail.com
> <mailto:richardhightower_at_gmail.com>>> wrote:
>
> Well said. Not sure if the async Servlets is the same beast
> though. Isn't that dealing with Comet type stuff?
>
>
> On Tue, Sep 27, 2011 at 11:47 AM, Pete Muir
> <pmuir_at_bleepbleep.org.uk <mailto:pmuir_at_bleepbleep.org.uk>
> <mailto:pmuir_at_bleepbleep.org.uk
> <mailto:pmuir_at_bleepbleep.org.uk>>> wrote:
>
> This, for me, makes it even more important. There is
> nothing a
> developer hates more than learning 3 ways to do the
> same thing!
>
> On 27 Sep 2011, at 18:30, Marina Vatkina wrote:
>
> > I'm not sure it'll be straight forward to add support for
> all managed beans - servlets and Web services already
> define
> their own approach to an asynchronous processing...
> >
> > -marina
> >
> > Antonio Goncalves wrote:
> >> Like Pete I would like to see @Asynchronous in
> managed beans.
> >>
> >>
> >> On Tue, Sep 27, 2011 at 12:46, Pete Muir
> <pmuir_at_bleepbleep.org.uk
> <mailto:pmuir_at_bleepbleep.org.uk>
> <mailto:pmuir_at_bleepbleep.org.uk <mailto:pmuir_at_bleepbleep.org.uk>>
> <mailto:pmuir_at_bleepbleep.org.uk
> <mailto:pmuir_at_bleepbleep.org.uk>
> <mailto:pmuir_at_bleepbleep.org.uk
> <mailto:pmuir_at_bleepbleep.org.uk>>>> wrote:
> >>
> >> I would like to see this too. However I see this
> only as
> a first
> >> step and would also like to add this to the list of
> services to
> >> extract entirely so they can be used in managed beans
> (similar to
> >> interceptors and our discussions around declarative
> transactions).
> >>
> >> Pete
> >>
> >> On 27 Sep 2011, at 02:19, Marina Vatkina wrote:
> >>
> >> > This was also voted for by 2 experts. Would anybody
> else like to
> >> see this change in EJB 3.2?
> >> >
> >> > thanks,
> >> > -marina
> >>
> >>
> >>
> >>
> >> --
> >> Antonio Goncalves
> >> Software architect and Java Champion
> >>
> >> Web site <http://www.antoniogoncalves.org> | Twitter
> <http://twitter.com/agoncal> | Blog
> <http://feeds.feedburner.com/AntonioGoncalves> | LinkedIn
> <http://www.linkedin.com/in/agoncal> | Paris JUG
> <http://www.parisjug.org>
>
>
>
>
> -- *Rick Hightower*
> (415) 968-9037 <tel:%28415%29%20968-9037>
> <tel:%28415%29%20968-9037>
> Profile <http://www.google.com/profiles/RichardHightower>
>
>
>
> --
> Antonio Goncalves
> Software architect and Java Champion
>
> Web site <http://www.antoniogoncalves.org> | Twitter
> <http://twitter.com/agoncal> | Blog
> <http://feeds.feedburner.com/AntonioGoncalves> | LinkedIn
> <http://www.linkedin.com/in/agoncal> | Paris JUG
> <http://www.parisjug.org>
>
>
>
>
> --
> Antonio Goncalves
> Software architect and Java Champion
>
> Web site <http://www.antoniogoncalves.org> | Twitter
> <http://twitter.com/agoncal> | Blog
> <http://feeds.feedburner.com/AntonioGoncalves> | LinkedIn
> <http://www.linkedin.com/in/agoncal> | Paris JUG <http://www.parisjug.org>