mikebridge commented on PR #44514: URL: https://github.com/apache/superset/pull/44514#issuecomment-5823503739
@rusackas Thanks for the careful review. I addressed the four points in [07a27cb](https://github.com/apache/superset/commit/07a27cb12d8d9389ff32609939482e68e5f138c5): - The 7.0 `SUPERSET_VERSION_HISTORY_RETENTION_DAYS` key remains a deprecated fallback for environment and Python config. Existing 0/negative disables and longer windows are preserved. A star-imported new default cannot silently shorten an old window; an ambiguous pair keeps the safer disable or longer window. - The upgrade notes warn that copying an old `-1` to the new key would change disable into immediate eligibility. Explicit new `-1` still works and logs a startup warning; use new `0` to keep pruning disabled. Invalid or oversized values defer cleanup. - Continuum's connection-scoped unit of work now holds one capture decision through commit/rollback, so a helper session cannot clear captured operations or re-enable a denied unit. Tests cover both directions. The host still must keep its predicate stable for the whole transaction, as documented. - `UPDATING.md` now says a parent/child snapshot rebuild failure aborts the save and requires rollback. The exact tree passed 578 affected unit tests, five retention integration tests on migrated SQLite, and all applicable hooks on the eight follow-up files. The full PR hook run in this isolated checkout hit only unbuilt frontend declaration artifacts (TS6305); fresh exact-head CI is queued. Shell #5218 remains dependent on this core change reaching a private production commit and a verified shell pin; this PR head is not that production SHA. Would you take another look when CI settles? -- 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]
