[grepjava]

DescribeJournal.decorateRouteToProperties printed each queue ID in a message's _AMQ_ROUTE_TO properties without its queue name, though the queue names were in BrokerInfo. References and acks are decorated through decorateQueueName; the route-to IDs were not, so DescribeJournalTest.testDecorateRouteToSingleQueue, testDecorateRouteToMultipleQueues and testDecorateRouteToClusterProperty fail on main, and every build since 4d82196 stops at artemis-server.

Each route-to ID is now followed by decorateQueueName(id, brokerInfo), which adds (queueName), or nothing for an ID that is not a known queue.

DescribeJournalTest passes, 6 of 6. It fails without the fix.

https://issues.apache.org/jira/browse/ARTEMIS-6242

[clebertsuconic]

What happened was.. I recently added decorations on upstream/main. Domenico pointed me that the HD_ROUTE_TO should not be decorated as these queueIDs belong to another broker (another journal).

So I removed the decoration, but had forgotten to remove / adapt the tests (which were actually generated by AI).

I am fixing the tests / removing them.

As of proper decoaring the HDR_ROUTE_TO, we would need to load the bindings of the target node... we could do that with an optional feature (like provide the bindings of a remote node), but I think that's an overkill for a feature I would personally like to have replaced by federation / amqp federation.

Thanks for the PR though.

[clebertsuconic]

I sent this PR to address this: #6843

[grepjava]

Thanks for explaining. I hadn't noticed the route-to IDs come from the other broker's journal (remoteQueueID), so looking them up locally would label them with the wrong queues. #6843 is the right fix.