add TCK tests for eager initialization (@Eager)

[Ladicek]

TCK tests for jakartaee/cdi#911

[Ladicek]

This is a draft, because:

• the spec PR (add @Eager cdi#911) wasn't merged yet
• there's a SNAPSHOT dependency for now

Other than that, this should be complete I believe.

[Ladicek]

Added tests for @Eager on a stereotype.

Also added a 2nd commit with missing tests for Bean.isReserve() -- I forgot to include those in a previous PR. I'm too lazy to open an extra PR for this :-)

[Ladicek]

Rebased and updated to reflect the latest spec text. Ready for review, although there is still a dependency on SNAPSHOT, so not ready for merging yet.

[Ladicek]

Rebased. No SNAPSHOT dependency anymore.

[manovotn]

Just a few small notes. Otherwise looks good

[Ladicek]

@manovotn See the last comments from @Azquelt -- I hope the one change of scope is OK for you, but I vaguely recall discovery works differently in Weld, so would be good to have another look.

[manovotn]

@manovotn See the last comments from @Azquelt -- I hope the one change of scope is OK for you, but I vaguely recall discovery works differently in Weld, so would be good to have another look.

Yea, that breaks my current implementation. Let me take a closer look at what's going on there...

[manovotn]

@manovotn See the last comments from @Azquelt -- I hope the one change of scope is OK for you, but I vaguely recall discovery works differently in Weld, so would be good to have another look.

Yea, that breaks my current implementation. Let me take a closer look at what's going on there...

@Ladicek the discovery is not the issue here, it's the validation.
IIRC, Weld does validation per-bean, not on the stereotype itself. And the bean it perfectly valid.

I think we had a similar discussion regarding alternative and reserve on a stereotype and I can see I added

[manovotn]

@manovotn See the last comments from @Azquelt -- I hope the one change of scope is OK for you, but I vaguely recall discovery works differently in Weld, so would be good to have another look.

Yea, that breaks my current implementation. Let me take a closer look at what's going on there...

@Ladicek the discovery is not the issue here, it's the validation. IIRC, Weld does validation per-bean, not on the stereotype itself. And the bean it perfectly valid.

I think we had a similar discussion regarding alternative and reserve on a stereotype and I can see I added

OK, fixed now, thanks for pointing it out 👍