users@servlet-spec.java.net

[servlet-spec users] Re: Clarification on CDI-452

From: Edward Burns <edward.burns_at_oracle.com>
Date: Tue, 11 Oct 2016 13:37:32 -0700

>>>>> On Wed, 05 Oct 2016 00:40:58 +0000, "John D. Ament" <john.d.ament_at_gmail.com> said:

JA> I'm looking for some kind of clarification on CDI-452, preferably from the
JA> Servlet expert group.

JA> The CDI EG is implementing request context management, which would
JA> allow application developers or framework developers to
JA> activate/deactivate request contexts.

JA> We have this ticket https://issues.jboss.org/browse/CDI-452 reported by Ed
JA> Burns, I would assume on behalf of the Servlet EG (at least at some level).

Yes, the Servlet spec has an interest in seeing that this works.

JA> Now I would say that we have this answered.

That's great news.

JA> However, the implication I'm seeing in the ticket is that you're
JA> potentially expecting the beans available in the original request to
JA> be available in the new async thread.

Looking at the code example, I think it would be acceptable if

@Inject
Bean bean;

which happens in the async thread, is only availbale in that async
thread, and not in the original thread. In other words, I would be ok
if there was some rule that says you can use it in the original thread,
or in the async thread, but not both.

JA> So I was wondering, is that actually expected? Or is it enough to say that
JA> a new request context is available within that thread?

I don't know enough about the implications of just making a new request
context available.

Thanks for re-engaging on this.

Ed

-- 
| edward.burns_at_oracle.com | office: +1 407 458 0017