Raw Bean Types

[dblevins]

The accepted test challenge jakartaee/cdi-tck#429 has identified some ambiguity with how raw types in should be handled in bean definitions.

More discussion is required to determine:

• What, if any, portability is offered here from a CDI specification perspective
• What, if any, updates to the TCK need to occur
• What, if any, updates to the specification need to occur to reflect the outcome

The factor that lead to the test's exclusion is behavior flagged by the OpenJDK team as a bug (JDK-8044366). The bug identifies two options, direct and indirect, for determining the supertypes of a generic type. The author notes javac currently implements the indirect approach and advocates this is a bug and the correct approach is direct. This perspective is not yet accepted and the issue will not be resolved at the JDK level till other related issues are resolved.

[manovotn]

Given the follow up discussion in the linked issue, I think the ambiguity was clarified as a misunderstanding.
We should be good to close this issue, agreed @dblevins? Or is there something else that this issue should cover?

As for the JDK bug, we agrees to accept both interpretations and we cannot do much until the JDK issue gets resolved (if ever).