no-hup commented on PR #44017: URL: https://github.com/apache/superset/pull/44017#issuecomment-5791105926
Thanks, both of these are fair. - **The staleness reasoning.** You're right that the "nothing reads a user's subjects after changing them" wording is wrong — `commands/utils.py:127` creates the subject and line 133 reads it straight back, and the group backfill does the same. I rewrote the docstring to the reason that actually holds: a subject created mid-request isn't listed in any resource's editors or viewers yet, so a check against a cache that omits it returns the same access decision. The window is bounded by the request either way. - **The test.** Added `test_get_user_subject_ids_serves_a_stale_set_within_the_request`: warm the cache with `[7]`, swap the query to `[7, 99]` in the same request, assert the call still returns `[7]`. That pins the staleness window as a contract instead of a docstring claim. - **The description.** Refreshed it so it says the cache lives on `flask.request`, not `g`, with the corrected staleness note, so the squashed message matches the code. The red checks are the repo-wide `check_pot_drift` / `babel-extract` failure you already confirmed isn't from this diff. -- 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]
