DanielLeens commented on PR #12922:
URL: https://github.com/apache/seatunnel/pull/12922#issuecomment-6557460762

   Thanks @fatmanverse and @SEZ9. Your new per-point reply matches the source 
pointers I checked on the unchanged head:
   
   1. `JobInfoService.java:504-523` first materializes 
`validatedSubmitJobsParams` for **all** entries, then enters the submission 
loop. This is parameter prevalidation, not an atomic transaction for later 
config/cluster failures.
   2. `RestUtil.java:103-107` rejects savepoint start with missing/blank 
`jobId`; `JobInfoService.java:556-575` rejects blank restore-source IDs and 
parses nonblank IDs with `Long.parseLong`. The existing messages name `jobId` / 
`restoreSourceJobId`; they do not use the literal `params.jobId` prefix. For a 
malformed JSON number, the earlier error deliberately names only params/the 
numeric shape, avoiding arbitrary client keys; I do not need another 
naming-only patch.
   3. `RestUtilSubmitJobsParamsTest.java:78-105,150` covers null skip, 
preserved empty strings, blank savepoint IDs and nested values. 
`JobInfoServiceSubmitJobsTest.java:107-162` puts an invalid entry **second** 
and asserts no submissions, including nested objects, fractional IDs, malformed 
string IDs and blank restore IDs. The HTTP companion is `RestApiIT.java:1450`.
   4. The EN/ZH v1/v2 submit-jobs parameter tables include string/number IDs, 
boolean flags, `restoreMode` and `restoreSourceJobId`; the v2 English contract 
is at `rest-api-v2.md:1138-1141`.
   5. `RestUtil.java:129-168` omits the client key and value entirely from type 
errors. This avoids the logging problem rather than trying to sanitize/truncate 
arbitrary keys.
   6. The JSON helper is `buildRequestParamsFromJson` (`RestUtil.java:93`); the 
URI helper remains `buildRequestParams(Map, String)`, so the ambiguous Map 
overload is gone.
   
   The description-only update does not change my source-clear conclusion. 
Current-head Build is green. The author's reported local test counts are their 
execution evidence, not tests I ran; a write-capable maintainer still owns 
final approval.
   
   Checked 2026-11-01T11:42:18.258131+00:00; unchanged head 
`74235a4c2272254277b80fd337ecb37df8ad2fa8`, base `dev`; compare `diverged`, 
ahead 10 / behind 10; merge `blocked` / `BLOCKED`; review decision 
`REVIEW_REQUIRED`. Behind/diverged is recorded independently and does not 
itself request a sync. Daniel has READ repository permission.
   


-- 
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