[cdi-dev] [servlet-spec users] [jsr369-experts] [116-CDIRelatedBeansInServletSpec] PROPOSAL

Edward Burns edward.burns at oracle.com
Fri Nov 21 09:14:25 EST 2014


Mark added cdi-dev to the recipients list, but I am not subscribed to
that (please don't add me, I rely on JJ to keep me up to date!), so this
may bounce there.

Also, Mark, there is no <jsr369-users at servlet-spec.java.net>.  The users
lists are re-used from JSR to JSR for each technology.  Thus, I have put
users at servlet-spec.java.net in place of the non-existent list.

On Wednesday, 19 November 2014, 19:40, Edward Burns <edward.burns at oracle.com> wrote:

[...]

EB> section 15.5.15 "Contexts and Dependency Injection for Java EE
EB> requirements".

EB> The CDI spec is trying to trim itself down [...]

EB> CDI Spec section 3.8 [1] places some requirements on CDI implementations
EB> when running with Servlet. To better suit user desires for modularity
EB> these requirements are better met by moving them to the Servlet
EB> spec. 

[...]

EB> To put these items in the Servlet spec, we may have to couch them in
EB> terms similar to, "when running in an environment that also supports
EB> CDI...".

EB> PROPOSAL:

EB> Include language in our spec section 15.5.15 which allows the CDI spec
EB> to remove their language while preserving desired existing user
EB> functionality.

[...]

>>>>> On Thu, 20 Nov 2014 13:47:19 +0000 (UTC), Mark Struberg <struberg at yahoo.de> said:

MS> I'm not quite sure that it's worth it. The @RequestScoped and
MS> @SessionScoped are already in the CDI spec package. And as you know
MS> we must not split packages between different specs.

>>>>> On Thu, 20 Nov 2014 22:32:39 +0100, arjan tijms <arjan.tijms at gmail.com> said:

AT> I don't think it's about moving those scope annotations, but just
AT> about the injection of HttpServletRequest, HttpSession and
AT> ServletContext.

Granted.  The proposal does not touch those.  Mark was just questioning
if it was worth it.  I am skeptical myself, but am responsive to
community input.

MS> It would feel quite unnatural to have annotations in the CDI APIs
MS> but the description in the servlet spec. [...]

I'll trust that Antoine Sabot-Durand (ASD) will take your reservations
into account when deciding to press on with CDI-492-FobStuffToServlet or
not.

MS> What we probably could add (regardless whether on CDI or Servlet
MS> side) is kind of a @WebAppScoped. This is really missing atm. But
MS> again this must also be active in e.g. JMS invocations which form
MS> kind a 'logical app'.

That's a whole new thing, again for ASD to consider.  As you observe, it
doesn't belong in Servlet because JMS can also use it.

Ed

-- 
| edward.burns at oracle.com | office: +1 407 458 0017


More information about the cdi-dev mailing list