SEZ9 commented on issue #12266: URL: https://github.com/apache/seatunnel/issues/12266#issuecomment-5902782948
Thanks @Rangsh — that summary matches the gate exactly, and nothing in it needs correcting. Confirming the points as you stated them: - The 4 GiB JMH fork budget and the `34317023975` full-WAL reload OOM on `432bdb3d9` are recorded as supporting acceptance context only. The `-Xmx2g` single-key smoke on `5bbc304b2` and any heap budget by itself are not evidence of selective loading or no-retention. - **Phase A** is the functional proof: deterministic `FileMapStore` / WAL-reader tests covering requested keys among filler keys, overwrite and tombstone cases, no unrelated keys returned, and no whole-map retention, with the constrained-heap diagnostic on the `initialStoredJobCount=1000` shape as like-for-like supporting evidence. Not relying on the SPI default full-load-and-filter path is the right call. - **Phase B** opens only after Phase A is merged and shown selective, widens the durability sample outside the measured `SingleShot` path, includes the tear-down-only assertion in its first revision, and references the merged Phase A change. On #12173: understood that the head stays at `7948e2640` with no further pushes unless a re-review asks for changes, and that the remaining gate is the stale docs `CHANGES_REQUESTED` already addressed in `9a6077c58` plus clearing the blocked state. Holding both phases design-only and not stacking on that branch until it is healthy and merged is correct. Remaining asks from my side: 1. When #12173 is merged, open Phase A and link it back here so the deterministic test coverage can be reviewed against the contract above. 2. In the Phase A description, state the heap setting used for the constrained-heap diagnostic and label it explicitly as supporting evidence, so it is not read as the acceptance criterion. 3. Do not open Phase B until Phase A is merged; when you do, point to the merged Phase A change in the PR description. No label or assignee changes from this side either. <!-- streview-comment:1413 --> -- 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]
