[
https://issues.apache.org/jira/browse/FLINK-40172?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18124834#comment-18124834
]
Weiqing Yang commented on FLINK-40172:
--------------------------------------
[~martijnvisser] [ |https://github.com/MartijnVisser] Thanks for catching this.
Reproduced on clean master, and the failing test is ours.
FLINK-38569 stopped rendering {{metadata=[]}} in scan digests and updated every
golden, including 28 lines of {{{}EarlyFireJoinHintTest.xml{}}}. This PR merged
about three hours earlier and added a new case to that same file, so the sweep
ran against a base where {{testEarlyFireProcTimeOnRowTimeJoin}} didn't exist
yet and its two {{TableSourceScan}} lines kept the empty spec. Both were green
on their own base.
The fix is a 2-line golden update, the same edit the sweep made to the 8
sibling cases. It already landed on master as
[{{b68f22c}}|https://github.com/apache/flink/commit/b68f22c91297ba5949780296a7f4e8ff2e484c53],
thanks @ [~snuyanzin] [nzin|https://github.com/snuyanzin].
> Support processing-time early fire on a row-time interval join
> --------------------------------------------------------------
>
> Key: FLINK-40172
> URL: https://issues.apache.org/jira/browse/FLINK-40172
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / Runtime
> Reporter: Weiqing Yang
> Assignee: Weiqing Yang
> Priority: Major
> Labels: pull-request-available
> Fix For: 2.4.0
>
>
> Add the domain-aware processing-time early-fire timer path so an event-time
> join can early-fire on wall-clock time; drop the temporary rejection;
> snapshot/restore replay coverage.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
