Should JWT @Claim require @Inject too

[emmanuelbernard]

The guide seems to use @Claim and @Inject together and 0.27 seems to mandate that. But our quarkus lab used to work with just @Claim in seems
Is that a regression or was the behavior never intended?

@lordofthejars

[sberyozkin]

This test shows Inject everywhere.
May be this code is related ? Not 100% sure, CC Martin @mkouba and Michal @michalszynkiewicz

[mkouba]

@Inject is optional for a field injection point if it has at least one qualifier. It's a Quarkus-specific feature that actually violates the CDI spec (where @Inject is mandatory).

@sberyozkin the TCK must be CDI-compliant. So it has to use @Inject.

In any case, I wouldn't call it a "regression"...

[sberyozkin]

@mkouba Hi Martin, thanks for the fast update; As far as I'm concerned, the fewer annotations, the better :-), so if Qurakus allows for it then it is fine. May be some documentation has to be updated somewhere to answer a question from @emmanuelbernard ? If you tell me what to update then I can take care of it, thanks

[mkouba]

I've updated the CDI guide to include few lines about this feature: #5167 (previously it was only mentioned in the blogpost)

[emmanuelbernard]

@sberyozkin can you run some quarkus specific tests as well outside the TCK. Because if I did not make a mistake it was not working for me at the time. I had to add @Inject. But maybe I got it wrong.

[sberyozkin]

@emmanuelbernard
Stuart updated this test resource recently, the test itself is here. So it works without Inject.

[mkouba]

Thanks @sberyozkin.