rusackas commented on PR #39278: URL: https://github.com/apache/superset/pull/39278#issuecomment-5188383422
Update: tested the `model_class=` theory directly (apache/superset#42789, draft, closed now). `db.Model is FABModel` came back `True` locally, confirming the registries do get unified, but the same CI failures happened anyway, so that wasn't the real cause. Pulled the actual failure logs this time instead of guessing from the summary. The real symptom is sharper than "NoneType has no attribute X" suggested: `db.session.bind` is `None` outright at the point of failure (`db.session.bind.dialect.name` blowing up in `soft_delete_tests.py`), and separately a `db.session.query(Dashboard).filter_by(slug="births").first()` right after a fixture set the dashboard up returns `None`. That's a session with no resolved engine, or one that can't see data a different context wrote. Circles back to FSA 3.0's actual changelog line, "the session is scoped to the current app context instead of the thread," more directly than the registry theory did. Sorry for the runaround. Not going to keep guessing in public, will dig into where `db.session.bind` resolution actually breaks down before saying more. -- 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]
