nsivabalan commented on PR #19206:
URL: https://github.com/apache/hudi/pull/19206#issuecomment-5222323227

   **Correcting an earlier note: the tests I said could not run locally do run, 
and they pass.**
   
   I reported a chain of "pre-existing local build failures" — `hudi-utilities` 
not compiling on a `KafkaAvroSchemaDeserializer` error, a stale 
`hudi-utilities-bundle` shadowing `hudi-common`, and `hudi-cli` therefore 
unable to resolve `MetaFieldsMode.parse`. I attributed all of it to the repo.
   
   It was my build invocation. The root pom defaults `kafka.version` to 
`2.0.0`, whose `Deserializer` interface lacks the `deserialize(String, Headers, 
byte[])` overload that `KafkaAvroSchemaDeserializer` declares with `@Override`. 
The `spark3` profile sets `kafka.version=3.4.1`, which has it. Everything 
builds with `-Pspark3` — and the bundle was stale precisely *because* 
rebuilding it hit that same error.
   
   ### What now actually runs
   
   | Suite | Result |
   |---|---|
   | `TestTableCommand` (#19206) | **19/21** — all 13 meta-fields-mode tests 
pass, including the 20-pair migration matrix and the sibling-direction test |
   | `TestHoodieStreamerMetaFieldsMode` (#19205) | **6/6** — including the 
restart test and the restart-stating-the-legacy-boolean rejection |
   
   Both of those had been committed with a note saying they were compile- and 
checkstyle-verified only. They are now genuinely executed. The streamer restart 
rejection was the one I was least comfortable shipping unverified, since it is 
the breaking change in the release notes.
   
   ### The two remaining `TestTableCommand` failures are genuinely pre-existing
   
   `testDefaultCreate` and `testFetchTableSchema` assert an archive path of 
`.hoodie/history` where the code produces `.hoodie/timeline/history` — a 
timeline-layout change. The assertion is on master's own copy of the file, last 
touched by unrelated commit `5124ad76b17d`, and neither PR edits that test. 
Confirmed by inspection rather than assumed.
   
   Apologies for the noise — the earlier claim was wrong, and it would have led 
anyone reproducing locally down the same dead end.
   


-- 
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