Caideyipi commented on PR #19241:
URL: https://github.com/apache/iotdb/pull/19241#issuecomment-6819286415

   ### Independent re-review of the implemented draft
   
   Completed one independent re-review of the full 13-file PR diff and relevant 
context at `a730680121d64c526ea2c41efc8961f9b6eced6b`, excluding the TODOs 
explicitly documented in the PR description. **No new actionable 
source-behavior or test-logic findings were confirmed within that scope; no 
fixes or branch updates were needed.** This is not approval of the unfinished 
recovery protocol.
   
   The review covered retryable versus permanent PREPARE/COMMIT failures, 
persisted COMMIT/ABORT identity and late replicated commands, snapshot 
selection/copy boundaries, legacy migration failures, cleaner/retention 
ordering, lazy manager deletion fencing, lock ordering, and regression 
assertions. In particular, the earlier fixes remain intact:
   
   - Preserved legacy COMMIT/ABORT decisions are included in snapshot history 
under their task IDs when migration cannot publish, with and without an 
initialized manager.
   - The permanent COMMIT exception test actually injects the COMMIT exception 
and checks the returned message.
   - A readable legacy decision remains authoritative when migration fails. The 
failure leaves staging intact and does not cache an unsuccessful migration, 
allowing later terminal lookups to retry publication.
   
   Validation: this round passed `git diff --check` for the full PR diff and 
confirmed a clean review worktree. No Maven tests were rerun because this round 
changed no source or tests. The previously reported evidence remains 
revision-specific: at `a730680`, the clean focused 28-module reactor ran 
`LoadTsFileManagerTest` with **40 tests passed, 0 failures/errors/skips**, 
recompiled datanode main and test sources, and reused 42 reactor goals from the 
Maven build cache; Spotless, Checkstyle, forbidden-API checks, and `git diff 
--check` passed. The earlier 81-test focused run and 15 added tests were run at 
`aa6b6acd`, not rerun here. All added tests have previously been run locally. 
ClusterIT, Chinese-locale compilation, process-kill recovery testing, 
performance testing, and full CI were not evaluated in this round; no CI was 
awaited or polled.
   
   The PR and source branch were refreshed and reconfirmed. The recorded PR 
base is `54bff3d9a421f9ba148fda3a26626df570422f54`; the merge base of the 
reviewed diff is `6e7448d9bebd644259b695dfc8774d3d9f91252e`; refreshed target 
`master` is `d47472156d793fdaa13b8f87051f363222eddd50`. Integration with the 
newer target was not built or tested.
   
   - Reviewed SHA: `a730680121d64c526ea2c41efc8961f9b6eced6b`.
   - Final local SHA: `a730680121d64c526ea2c41efc8961f9b6eced6b`.
   - Confirmed source-branch SHA: `a730680121d64c526ea2c41efc8961f9b6eced6b`.
   - Confirmed PR HEAD: `a730680121d64c526ea2c41efc8961f9b6eced6b`.
   
   **The PR remains draft.** The documented TODOs are still incomplete and were 
neither expanded nor treated as resolved by this 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]

Reply via email to