yujun777 commented on PR #68390: URL: https://github.com/apache/doris/pull/68390#issuecomment-5810708128
Answering the summary review at `6c6f79d5403`. Both threads it names as still unresolved are addressed in 8934d283cc4; the third is left open on purpose. - The strict-`PARTITIONS` refusal being unreachable behind the initial-refresh shortcut was this PR's hole -- the two rules answer COMPLETE and the one that refuses ran second. Judged first now, with a unit test that asserts both that the shortcut is active in that state and that the request is refused. - The generic `processBaseTableChange -> Env.alterMTMVStatus` path applied the status under the lock and enqueued after releasing it. Fixed, though the path predates this PR (`git blame`: 2023-11-24): it now goes through `MTMV.invalidateWholeMv`, the shape this PR introduced and the one the sibling invalidation in the same loop already uses, and `Env.alterMTMVStatus` is deleted. Same state, same detail, byte-identical payload, replay untouched; the DDL now awaits the record, which those sibling branches already do. - STOP's `cancel(false)` is left open with the reasoning on the thread: it is the job framework's cancel semantics, shared by every job type, and the accumulator-publication half the finding named is fixed. The PR description is refreshed: the `refreshEpoch == 0` exemption and the "a rename no longer moves it out of `NORMAL`" claim are gone, the behaviour-changed list and the release note carry the routing changes, and the test section names the new suite. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
