The GitHub Actions job "Tests (AMD)" on 
airflow.git/feature/make-basedagbundle-logging-dry has failed.
Run started by GitHub user dabla (triggered by dabla).

Head commit for run:
1322642e8e8b40c2e72afe9b363dc4269374c3f7 / David Blain <[email protected]>
Fix _log setter, naming, types and test isolation in BaseDagBundle

Released bundle providers (git, amazon, google) assign self._log in
__init__ against apache-airflow>=3.0.0 with no upper bound. Without a
setter the read-only property raises AttributeError on core upgrade,
breaking every installed bundle. Adding a setter keeps the lazy default
while accepting direct assignment.

Rename _log_context to _get_log_context so the method reads as a
callable rather than an attribute (AGENTS.md naming rule). Tighten
annotations: dict[str, Any] return type, -> Any on the property, and
an explicit Any type on the backing __log attribute.

Add tests in test_base.py that cover the lazy cache, bundle_name/version
bindings, subclass _get_log_context merging, and the setter — the
subclass context test is the only one that fails without the PR's change.

In the compat test suite: move function-level imports to module scope,
extract a legacy_compat_base fixture that reloads the module inside the
patch and reloads again on teardown so sys.modules is never left in the
fallback state, and replace the three identical _DummyBundle definitions
with a _make_bundle_class factory.

Co-authored-by: Copilot <[email protected]>

Report URL: https://github.com/apache/airflow/actions/runs/29522660497

With regards,
GitHub Actions via GitBox


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to