Hi All, To clarify my intent behind #590 [1], the main idea behind this refactoring is to establish something akin to a TCK [2] for Polaris.
These tests are meant to be used to validate that an alternative server runtime (e.g. Quarkus) behaves according to what is expected from the Polaris Server. Hopefully all implementations that pass these test will be mutually interchangeable (but we may need to add more tests to reach that goal). Cheers, Dmitri. [1] https://github.com/apache/polaris/pull/590 [2] https://en.wikipedia.org/wiki/Technology_Compatibility_Kit On Tue, Dec 31, 2024 at 8:32 AM Alex Dutra <alex.du...@dremio.com.invalid> wrote: > Hi all, > > I hope everyone is having a great holiday season so far! > > Unless I'm mistaken, I think we have an agreement to move forward with the > switch to Quarkus. I'm wondering when we should merge it? Keeping such a > large > PR open for too long is not ideal, but I also don't want to rush it. > > Here 's some useful context to help us decide: > > The PR has been up for review for a few weeks now, and I encourage > everyone to > take a look at it: : https://github.com/apache/polaris/pull/469. > > See also the proposal document: > > > https://docs.google.com/document/d/1C7E0ma6OvHGvWMJlZVkQHBsvPg8Pe9aroDtQ_79SIXY/edit > > Luckily, the changes are not too invasive for most modules – except for the > Dropwizard runtime module, of course. People can continue working on > modules > like core, api, persistence, service-common, etc. without much trouble. > > OTOH, merging soon-ish would settle the dust, and we could start focusing > more > on the new features that we have planned for the next releases. > > Merging would also unlock a few other PRs that are waiting on the Quarkus > switch: > > - https://github.com/apache/polaris/pull/589 (follow-up to the Quarkus PR) > - https://github.com/apache/polaris/pull/592 (follow-up to the Quarkus PR) > - https://github.com/apache/polaris/pull/594 (fixes a sensitive issue) > > I looked briefly at all the outstanding PRs, and I think that the only one > that > might be affected significantly by the switch to Quarkus is the one that > refactors the integration tests: > > https://github.com/apache/polaris/pull/590 > > In fact, it might be a good idea to merge that one first, since it would > somewhat simplify the Quarkus PR itself. That's up for discussion, though. > It > could be merged after as well. > > Also, Mike's PR that refactors PolarisAuthorizerImpl would be affected, > albeit > in a lesser proportion, by the switch to Quarkus: > > https://github.com/apache/polaris/pull/465 > > I personally think that the changes in that PR are not too difficult to > merge > after the switch, and I could help with that. But again, that's up for > discussion. > > Merging the Quarkus PR won't be "the end of the road" for the Quarkus > switch. > There are still a few things that we need to do immediately after, like: > > - Adapting the documentation > - Adapting the Docker images > - Adapting the Helm chart > - Communicating the change to the community > > All of these are on my TODO list with high priority. > > And finally, I wanted to mention that the switch in overall aims to be as > seamless as possible for users and operators of Polaris, but a few things > will > change, and we need to be prepared for that: > > - The configuration file will have a new format > - Cross-cutting concerns like logging, metrics, and tracing will have > different > shapes, names, etc. > > These are noted in the proposal document, but I wanted to mention them > here as > well so that everyone is aware. > > I am hoping that the switch could be included in the next release, 0.9.0, > somewhere in January 2025. I'm looking forward to hearing your thoughts on > this. > > I wish you all a great New Year's Eve and a happy New Year! > > Best regards, > > Alex >
