CodeWithPravinMaske commented on PR #12608:
URL: https://github.com/apache/seatunnel/pull/12608#issuecomment-6187468143

   @SEZ9 @DanielLeens the branch is synced with the latest `dev`, which 
includes #12597 (`42d2aab3f`), so the merge-order condition is met. The only 
conflict was the two `incompatible-changes.md` entries (en/zh), now side by 
side.
   
   Pointers for the re-review, as asked:
   
   - **Parser fix, and removal of `restoreNullableColumns`**: commit 226cdde90. 
It adds `io/debezium/connector/mysql/antlr/MySqlAntlrDdlParser.java` (only 
`parsePrimaryIndexColumnNames` changes) and brings `MySqlSchema.java` back to 
`dev`, which is why that file is not in "Files changed". Commit fa70f3eb8 adds 
the override comments, the restore E2E and the `primaryKeys` test.
   - **Streaming DDL tests**: 
`MySqlSchemaTest#testNullableUniqueKeyColumnStaysOptionalInStreamingDdl` runs 
`ALTER TABLE ... ADD UNIQUE KEY` and `CREATE UNIQUE INDEX` through 
`MySqlDatabaseSchema#parseStreamingDdl` and asserts the column stays optional. 
`testRealPrimaryKeyColumnIsNotOptionalInStreamingDdl` is the control: a real 
primary key stays NOT NULL.
   - **Binlog DDL and checkpointed history, end to end**: 
`AbstractMysqlCDCITBase#testMysqlCdcNullInNullableUniqueKeyWithoutPrimaryKey` 
and `#testMysqlCdcNullInNullableUniqueKeyAfterRestore` both run `ALTER TABLE 
uk_added_null ADD UNIQUE KEY` during the binlog phase. The second one then 
savepoints and restores, and asserts that NULL (and no `0`) reaches the sink.
   - **`primaryKeys`**: 
`MySqlSchemaTest#testNullableColumnConfiguredAsPrimaryKeyEmitsNull`.
   - **Docs**: `docs/{en,zh}/connectors/source/MySQL-CDC.md` and 
`docs/{en,zh}/introduction/concepts/incompatible-changes.md`.
   
   One addition in the sync: now that both fixes are on the branch, #12597's 
`testMysqlCdcSnapshotWithNullableUniqueKey` compares complete rows (`id, code, 
name`) instead of `id, name`. So it also checks that the NULL codes reach the 
sink as NULL. This was @SEZ9's request on #12597. Locally, all three 
nullable-key E2E tests pass on Zeta and on the Flink containers that started. 
Two Flink images couldn't be built on my machine, because the JDK 11 image 
setup from #11545 failed there. CI will cover those containers.
   


-- 
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