kokhlo commented on PR #43960: URL: https://github.com/apache/superset/pull/43960#issuecomment-5598152051
Thanks @joe-clickhouse — really thorough analysis, much appreciated. Addressed in 009429a6: - **Fractional seconds**: added a `DateTime64(3)`-shaped case (`.123000+03:30` and short `.123` form) asserting the passthrough keeps milliseconds verbatim. - **DST fall-back pair**: added a `2026-11-01T01:30:00` pair with `-04:00`/`-05:00` asserting the parsed epochs stay one hour apart (and match the correct UTC instants) — the case you observed the old normalizer collapsing. - **`Fixes #43927` removed** from the description: you're right that the reporter's chart path never runs through `normalizeTimestamp` (epoch data mode), and the actual display-half regression is the backend epoch conversion from #37979; the string passthrough here is the SQL Lab / string-mode half. Left a note about the dataset-timezone feature path (#37014). The core claim stands per your verification: the string-form change is correct and worth keeping as regression coverage for the passthrough behavior. -- 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]
