SEPURI-SAI-KRISHNA commented on PR #11987:
URL: https://github.com/apache/seatunnel/pull/11987#issuecomment-5489169170
Thanks @DanielLeens, that is a genuinely useful catch and I have verified it
independently.
`AzureCosmosDBSourceSplitEnumerator.java:143` does contain a byte-for-byte
instance of the masking pattern this PR consolidates:
```java
private static int getSplitOwner(Integer splitId, int numReaders) {
return (splitId.hashCode() & Integer.MAX_VALUE) % numReaders;
}
```
I re-swept the current worktree at this PR's head to make sure nothing else
slipped in alongside it. `& Integer.MAX_VALUE) %` has exactly one production
hit, the Azure one above; `& Long.MAX_VALUE) %` has zero production hits;
`0x7FFFFFFF` has zero hits anywhere in the tree. So that line is now the only
remaining unmigrated instance, and your read is right that it arrived from
`dev` via #11167 rather than being something this PR skipped.
One extra detail worth folding into the follow-up: the parameter is already
an `Integer`, and `Integer.hashCode()` returns the value itself, so the
`.hashCode()` call there is an identity no-op. The migration is therefore not
just a swap but a small simplification:
```java
return HashUtils.bucketIndex(splitId, numReaders);
```
I will open that as a separate one-line PR once this one merges, rather than
growing this diff. Doing it now would invalidate your review and re-trigger the
full CI matrix for a change that is genuinely independent, and it also reads
more honestly in the history as "migrate the connector that landed after the
consolidation" than as a late addition here.
On the merge state: `reviewDecision` is `APPROVED` and `mergeable` is
`MERGEABLE`. `mergeStateStatus` is `UNSTABLE`, but I do not think that reflects
a failing check. All three apache-side check runs (`Build`, `Notify test
workflow`, `labeler`) are green on `dba3522ae`, and the legacy combined-status
endpoint returns `total_count: 0` for this head, which is the usual source of
an `UNSTABLE` reading on a commit that has check runs but no classic statuses.
For completeness, the fork run you dereferenced has since been rerun and is
fully green end to end: [`33364655791` attempt
2](https://github.com/SEPURI-SAI-KRISHNA/seatunnel/actions/runs/33364655791)
concluded `success` at 2026-08-31T17:59:48Z, 83 passed, 10 skipped, 0 failed.
The four failures on attempt 1 were three Maven Central transfer errors
(`junit-jupiter-params:5.11.0`, `seatunnel-shade-commons-lang3:3.18.0-3.0.0`,
`hive-llap-common:3.1.3:tests`) and a `doris-connector-it` container port
timeout, none of which reached thi
s PR's code.
The branch is now `ahead_by=2, behind_by=5` against `dev`, and none of those
five commits touch any file this PR changes. Happy to sync if a committer
prefers it, though I would otherwise leave the head where it is so the green
run above is the one that gets merged.
--
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]
