AyoubOm commented on PR #15607: URL: https://github.com/apache/kafka/pull/15607#issuecomment-2088456658
> Can we update existing test to not use `Map` but `List` (as I mentioned on the duplicate ticket https://issues.apache.org/jira/browse/KAFKA-16644) which should expose the bug? > > Need to think about the fix a little more (need to refresh my memory on the different "subscription types"). Ok, will check the integration test and update it accordingly. For the subscription types, I've kept a summary that helped me write the unit tests and implement the fix. It can make it easier to review it first if I missed something before checking the code ;) ``` // All cases of SubscriptionSendProcessor /* LEFT JOIN Tests Summary old record -> new record result Instruction (pk1, null) -> (pk1, null) -> NOTHING (pk1, null) -> (pk1, (null, ...)) -> PROPAGATE_IF_NO_FK_AVAILABLE (pk1, null) -> (pk1, (fk1, ...)) -> PROPAGATE_IF_NO_FK_AVAILABLE (pk1, (null, ..)) -> (pk1, null) -> PROPAGATE_IF_NO_FK_AVAILABLE (pk1, (null, ..)) -> (pk1, (null, ...)) -> PROPAGATE_IF_NO_FK_AVAILABLE (pk1, (null, ..)) -> (pk1, (fk1, ...)) -> PROPAGATE_IF_NO_FK_AVAILABLE (pk1, (fk1, ..)) -> (pk1, null) -> DELETE_AND_PROPAGATE (pk1, (fk1, ..)) -> (pk1, (null, ...)) -> DELETE_AND_PROPAGATE (pk1, (fk1, ..)) -> (pk1, (fk1, ...)) -> PROPAGATE_IF_NO_FK_AVAILABLE (pk1, (fk1, ..)) -> (pk1, (fk2, ...)) -> DELETE_NO_PROPAGATE + PROPAGATE_IF_NO_FK_AVAILABLE */ ``` I represent each record with (pk, (fk, ...)) where pk is the key, and the value is composed of the foreign key fk and the other fields represented by the dots (which do not impact the resulting instruction type) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: jira-unsubscr...@kafka.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org