Unfortunately there is kinda an overlap/conflict/mismatch (conflict is too strong, overlap
too weak ;-) between the Java EE spec and CDI here. CDI implies that the Servlet injection
point should work, however Java EE says it definitely won't. Assuming that Java EE
overrules CDI here, then you are correct, and that is how all containers I know of are
implemented.
On 3 Feb 2012, at 22:28, Mark Struberg wrote:
yes, that is exactly what should happen.
Please note that you could also have a WEB-INF/classes/META-INF/beans.xml to pickup
WEB-INF/classes as bean archive.
I always do it that way, because Eclipse and Idea can handle this much better when you
e.g. need to debug a webapp with mvn jetty:run
LieGrue,
strub
> ________________________________
> From: Joseph Snyder <j.j.snyder(a)oracle.com>
> To: cdi-dev(a)lists.jboss.org
> Sent: Friday, February 3, 2012 9:34 PM
> Subject: [cdi-dev] War with no WEB-INF/beans.xml
>
>
> I have a test war that does not contain WEB-INF/beans.xml. However, there is a jar
in WEB-INF/lib that does contain META-INF/beans.xml. According to the spec, 12.1, this
war contains 1 bean archive, the one in the jar. (If there were a WEB-INF/beans.xml then
there would be 2 bean archives in the war.) So when the war is deployed only the classes
in the jar are managed by the CDI container. Therefore, if there is an injection point
defined in a servlet (in WEB-INF/classes) then that injection point will never be
processed by the CDI container. Is my interpretation of the spec correct?
>
> _______________________________________________
> cdi-dev mailing list
> cdi-dev(a)lists.jboss.org
>
https://lists.jboss.org/mailman/listinfo/cdi-dev
>
>
>
_______________________________________________
cdi-dev mailing list
cdi-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/cdi-dev