Steven Hawkins

@vmuzikar @mabartos after looking more at #34095 - that is not something that is addressed by this PR. It is a separate, and pre-exisitng problem - there is only a...

> @shawkins By pre-existing problem you mean it's not a regression? Yes, the import logic has been that way for several releases.

> Yes, the import logic has been that way for several releases. Actually the issue is with the initial change of straightening out the transaction logic - in doing that...

> So it is a regression in 26. Yes

Moving to a draft - the user who opened the issue will contribute a pr https://github.com/keycloak/keycloak/issues/32348#issuecomment-2315515028 If for whatever reason that doesn't happen, I'll update this with the approch from...

> Sorry @shawkins, this PR got buried and I completely missed the notifications too :( No problem, I should have checked in on it. Didn't notice it still needed attention...

@brunobat https://github.com/keycloak/keycloak/blob/ecbd856176fa81203be0fe07464299d39883e9f8/operator/src/test/java/org/keycloak/operator/testsuite/integration/KeycloakDeploymentTest.java#L391 The only thing novel there vs the other operator tests are the Keycloak settings tracing-enabled=true (toggles quarkus.otel.traces.enabled to true) and log-level=io.opentelemetry:fine. There are similar tests running against an...

@brunobat the upgrade to 3.14.2 is still in progress: #32519 Once that is completed we'll be able to see if the issue reoccurs in github actions - it wasn't reproducible...

@brunobat the fix does not appear to have been successful. We have disabled the affected test: https://github.com/keycloak/keycloak/pull/33118/files#diff-66841197b0c15bb09bad83420bdbc5fcf04bc718225c5a3fb0dd10e4c7fa311fR393 cc @jonkoops @geoand

> This can be achieved by changing the @Target on @Pattern to include ElementType.TYPE_USE, and detecting the presence of the annotation on the type parameters in the declaration to include...