[cdi-dev] [TCK] Multi ClassLoader unit tests?

[Mark Struberg]

> To be honest I don't quite understand... could you describe your use-case it 
> in more detail?
> 
> So far it seems you have an EAR with one EJB jar "A" that contains 
> interface X. And two WARs "B" and "C" - both contain 
> @SessionScoped impl and declare dependency on "A".



not a dependency on A but they IMPLEMENT A.

There is no impl of A at all on the shared EJB level, just on webapp level.

Imo both Weld and OWB currently create their proxies based on the type of the injected Bean. But we must create the proxy based on the type of the injection point instead!


LieGrue,
strub



----- Original Message -----
> From: Martin Kouba <mkouba at redhat.com>
> To: cdi-dev at lists.jboss.org
> Cc: 
> Sent: Monday, June 4, 2012 9:30 AM
> Subject: Re: [cdi-dev] [TCK] Multi ClassLoader unit tests?
> 
> Hi Mark!,
> 
> Dne 3.6.2012 17:03, Mark Struberg napsal(a):
>>  Hi folks!
>> 
>>  I know that we did not address this area at all in the spec and thus have a 
> few problems in our implementations in this area.
>> 
>>  Where to look at if I like to contribute a unit test for a 
> multi-classloader scenario in the TCK?
> 
> 1. Do a pull request
> 
> Current development branch (CDI TCK 1.1) is master on GitHub repo:
> https://github.com/jboss/cdi-tck
> 
> Note that all tests are arquillian-based.
> 
> 2. File an issue
> 
> https://issues.jboss.org/browse/CDITCK
> 
>> 
>>  I like to add the following test:
>> 
>>  parent ClassLoader which has an interface X
>>  2 child ClassLoader which have @SessionScoped implementations A and B which 
> are independent of each other.
>>  Try to @Inject X x; into an EJB and reuse it in both child classloaders 
> ('webapps'). We could set to poolsize to 1.
> 
> To be honest I don't quite understand... could you describe your use-case it 
> in more detail?
> 
> So far it seems you have an EAR with one EJB jar "A" that contains 
> interface X. And two WARs "B" and "C" - both contain 
> @SessionScoped impl and declare dependency on "A".
> 
> Now what's the point?
> 
> 
>> 
>>  LieGrue,
>>  strub
>> 
>>  _______________________________________________
>>  cdi-dev mailing list
>>  cdi-dev at lists.jboss.org
>>  https://lists.jboss.org/mailman/listinfo/cdi-dev
> _______________________________________________
> cdi-dev mailing list
> cdi-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/cdi-dev
>