On Dec 11, 2008, at 8:58 PM, Vamsavardhana Reddy wrote:
Section 5.4.1 in SCA JEE Integration spec v1.0, the table at line 484 lists HttpSessionBindingListener. Do you think we should get rid of that entry?
Yes. IMO the only things that can plausibly have container mediated dependency injection are the objects the container creates.
On Thu, Dec 11, 2008 at 11:15 PM, David Jencks <firstname.lastname@example.org>
Why do you think that HttpSessionBindingListeners should support dependecy injection? They aren't managed objects AFAICT and as you state. Is there something I've overlooked in the spec?
On Dec 11, 2008, at 7:12 AM, Vamsavardhana Reddy wrote:
As part of SCA JavaEE Integration, I am looking at dependency injection into HttpSessionBindingListener's in web apps. HttpSessionBindingListener is different from other listeners in that HttpSessionBindingListener need not be specified in web.xml. HttpSessionBindingListener gets activated automatically when an object of that type (implementing HttpSessionBindingListener actually) is bound to HttpSession. Other listeners are specified in web.xml and their instantiation is done by the web container. Because of this, at deployment time, the web builder does not even know if there is any HttpSessionBindingListener that it should process to support dependency injection. Any thing I am failing to see?