Well, this is why there are passivation listeners and stuff in the Servlet spec.
We could easily also send a specific CDI event for it. But there is no such
event in the CDI spec so far.
The @Destryoed and @BeforeDestroyed are specifically for _destroyal_.
LieGrue,
strub
> Am 08.02.2018 um
Hmm, it is more vicious cause if the session is not destroyed you can still
want to trigger this event. Guess it is another case where both cases are
desirable (i want to clean up related state of the session...as well as I
don't want to touch the session)...
Since the appcontext destroy can be
Yea, it's mainly testing whether the @Observes
@BeforeDestroyed(SessionScoped.class) and @Destroyed(SessionScoped.class) do
work.
The tests itself are fine, but instead of relying that the sessions get
destroyed at server shutdown they could also have used Session.invalidate()...
LieGrue,
2018-02-08 11:28 GMT+01:00 Mark Struberg :
> All the embedded tests are now green.
>
> I'm now working on cdi-tomes (webprofile TCK).
> So far we have 10 errors, but a few TCK tests are broken because they
> wrongly assume that a container stop also kills the Session.
>
All the embedded tests are now green.
I'm now working on cdi-tomes (webprofile TCK).
So far we have 10 errors, but a few TCK tests are broken because they wrongly
assume that a container stop also kills the Session.
I've challenged those tests. Still have to review every red test...
LieGrue,