cansjt commented on issue #20055: URL: https://github.com/apache/airflow/issues/20055#issuecomment-986276727
As far as I am concerned if this is deliberate this is not a great choice. As mentioned this date comes from the output of the Parsing of a YAML file. For now we implemented a work around: intercept the output of the parsing before passing it to Airflow. But, to me, using pendulum is a choice internal to Airflow. At it's interface Airflow should be capable of accepting anything that comes from Python standard library. And should not impose to use anything extra to deal with something as simple as dates. To the least the DAG class should validate properly validate its input, as you put new constraints on the input it accepts. Also this change is not really [documented](https://airflow.apache.org/docs/apache-airflow/2.2.2/_api/airflow/models/dag/index.html#airflow.models.dag.DAG) (still states here `start_date` should be a `datetime.datetime`), and all the examples I could find accept `datetime` as input. And again it was working fine. Something that worked, but does no more, should fit anyone's definition of a bug (or so I hope). -- 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]
