hello-stephen opened a new pull request, #68800: URL: https://github.com/apache/doris/pull/68800
The load-channel profile regression receives named JSON objects from the profile-list API but indexes each row as an array. The resulting empty query ID only emits a warning, so a successful test can skip all profile assertions even when the INSERT profile is present. Read named fields and bind the profile to this run’s LOAD task, internal catalog, database, and unique table. Retry delayed profile availability and require a nonempty profile containing one of the original load-channel markers. Missing, malformed, or unrelated profiles cannot pass. Keep the original result assertion and clean up the unique table. Execute the credential-bearing INSERT through JdbcUtils on the same framework session, and omit raw SQL, profile responses, and potentially sensitive exception text from logs/errors. Validation: 23 behavior checks execute the production helper and suite fixtures, including named rows, unrelated/duplicate/missing IDs, delayed/empty profiles, both session types, and sanitized exceptions. Full suite syntax and diff checks pass; the fixtures also pass independently on JDK 8 with Groovy 4.0.19. No live regression rerun has been performed. The poll-start budget is 30 seconds, with at most 30 attempts. Each HTTP connect/read timeout is five seconds; an in-flight request can exceed the poll-start budget, and the read timeout is an inactivity limit, not a strict overall deadline. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
