SEZ9 commented on PR #12904: URL: https://github.com/apache/seatunnel/pull/12904#issuecomment-6593655371
Thanks for the follow-up on `fb542b3e9` — the two new cases in `test_polling_stops_when_the_job_is_not_known` are exactly what I was after for F2/F4: - `non-dict-body-then-marker` (2 polls) pins down that a single tolerated non-dict body no longer produces the permanent "job not known" verdict, while an explicit `found: false` still terminates promptly. - `status-less-good-status-less-then-marker` (4 polls) covers the `missing_response_seen` reset after a valid `jobStatus`, which is the part a regression would most likely break. Having the fake keep returning a status-less body once the scripted responses run out, so a regression trips the max-poll bound and the call-count assertion instead of leaking `StopIteration` into the catch-all, is a nice touch — the test fails for the right reason. On the recheck: not done yet on my side; I'll run the F1–F8 pass against `fb542b3e9`. Since you noted only `tests/test_diagnose_cli.py` changed in this push, two things would help me close it quickly: 1. For F1/F3 (docs for `found: false` on the other endpoints), F5 (the "The id may be wrong" wording), F6/F8 (UI fetch error path and showing the looked-up job id), and F7 (`job_exists` on non-dict payloads) — please confirm whether each is already addressed on this head or intentionally deferred, with a one-line pointer where applicable. 2. Your comment appears to have been cut off after "back to `REVIEW_REQUIRED` and" — if there was a remaining ask there, please repost it. Once I have that, I'll post the formal F1–F8 result. <!-- streview-comment:2232 --> -- 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]
