Gather bean validation violations in a single field error

[dpolysiou]

Change BeanValidationError path to go from root to field, in accordance with the spec.

[jmartisk]

@dpolysiou could you please rebase, specifically we need the branch to contain a076d7d (which I just merged a few minutes ago) to have the CI tests work properly

EDIT: now also 1dd98ec :)

[dpolysiou]

I saw one test in the wildfly-graphql repository that breaks (specifically, this one), because it has an assertion for the existing BeanValidationError extensions. It's a small change, but I don't know how the changes between these repositories should be coordinated.

[jmartisk]

If you're willing to help us fix the test too, then this is how we normally do it:

• I create a special branch in the wildfly-graphql-feature-pack repo named smallrye-graphql-2.11 or so
• That branch will capture all the changes that we plan to apply along with upgrading to SR-GQL 2.11 (so, fixing the test..)
• We temporarily change the SR-GQL CI here: https://github.com/smallrye/smallrye-graphql/blob/main/.github/workflows/build.yml#L61 to run against that temporary branch instead of main
• After the upgrade, we switch the CI back to wildfly-graphql-feature-pack's main

So if you want to help with that, I can create the smallrye-graphql-2.11 branch in wildfly-graphql-feature-pack and you can send a PR to it. But no problem if you want to let us handle it ourselves.

[jmartisk]

Ah, GraphQLBeanValidationTest in Quarkus will have the exact same problem. I already have a Quarkus branch (https://github.com/jmartisk/quarkus/tree/smallrye-graphql-2.11) where this would have to be fixed :)

Do you want me to handle the WildFly and Quarkus tests?

[dpolysiou]

Thank you, I will try to make a PR to these branches as soon as possible.

[jmartisk]

Thanks for all the PRs!
I've updated the CI to run against the temporary branches... Hopefully it will all pass :)

[jmartisk]

It worked, thanks a lot!
I'll take care of updating the wildlfly and quarkus repositories after releasing smallrye-graphql 2.11.

[schulzp]

@dpolysiou - This fails now, if the invalidValue is null, since Map.of does not allow null values