DanielLeens commented on PR #11746: URL: https://github.com/apache/seatunnel/pull/11746#issuecomment-6227359410
@li3zhi4 thanks for the careful evidence write-up, and a correction on my side first: you are right about the SQL Server ordering in my previous CI update (head `e7586c06a8`). I wrote that `JdbcSqlServerSplitIT` was "followed in the same JVM by `JdbcSqlServerIT`" and described that as the class-ordering case `alwaysPull()` protects. That is the reverse of the case that matters (IT teardown, then SplitIT). I re-read the fork run (38544126048) job logs just now: - SQL Server (17 lane, the only lane where it appears in the logs I read): `JdbcSqlServerSplitIT` ran first (2/2 passed), then `JdbcSqlServerIT` (10/10 passed). So this run does NOT exercise IT-teardown-then-SplitIT for SQL Server. Please treat that part of my earlier comment as retracted; the SQL Server re-pull evidence is the ordered local run you and @SEZ9 already discussed. - Oracle (both JDK lanes): `JdbcOracleIT` ran first (`Tests run: 39, Failures: 0, Errors: 0`), then `JdbcOracleSplitIT` (`Tests run: 3, Failures: 0, Errors: 0`, about 69-75 s). So for Oracle, CI on this head did execute exactly the IT-then-SplitIT order with real test methods and zero failures, which seems to be the evidence SEZ9 asked for. Your fresh local ordered run is still welcome as a cross-check, but this CI log already covers the order. I have not re-derived the F1-F8 items; those are for @SEZ9 to confirm as you requested. -- 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]
