unbridled-41 commented on PR #4737: URL: https://github.com/apache/rocketmq-dashboard/pull/4737#issuecomment-5761674699
@lizhimins Thanks for the detailed review — all points addressed in 1af1ed37 (branch rebased onto trunk `29a7d4ab`). **Both decisions, resolved:** 1. **Preview parity**: `preview()` now calls the same `requireExactSourceMessage` guard that `execute()` runs before publishing, so the agent can never present a plan it cannot carry out. The plan description also carries a warning documenting the refusal contract. Tests: `previewRefusesALossySourceInsteadOfPromisingAnUnexecutablePlanTest` and `previewAndExecuteUseTheSameExactSourceGuardTest`. 2. **Properties precision — narrowed**: you're right that the flag counts system keys that redelivery discards anyway. The guard now fires only on provable user-property loss: a visible user property whose value was abbreviated by the display cap. Cases fixed: a `propertiesTruncated` flag caused purely by system keys, and an intact user-property set beyond the display cap, both now redeliver; a truncated *user* value is still refused. `MessagePropertyDisplay` exposes its abbreviation marker so the guard can't drift from the projection. (Entries dropped entirely by the cap are unknowable from the projection, so the guard does not speculate about them.) **Docs**: `findMessageForRedelivery`'s javadoc no longer claims "exact source message" — it states the display-projection contract. The PR description now states plainly that `rmq.message.redelivery` has no REST/web caller, DLQ resend covers only `%DLQ%`, and ordinary-topic oversized/binary/over-cap messages therefore have no exact-redelivery route at all in the current architecture. **Tests after rebase**: `MetadataServiceTest` **61/61** (48 trunk + 13 PR, re-ran as requested with `-Dspring.profiles.active=dev`), handler tests **8/8**, `MessagePropertyDisplayTest` **8/8**; the new precision tests verified red against the previous guard (system-key-only flag refused → now proceeds). Ready for re-review. -- 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]
