Thank you everyone. My concerns have been addressed. I still think we should somehow advise users not to use this new trigger rule explicitly but that is something we can talk in the PR itself.
בתאריך יום ג׳, 28 במרץ 2023, 09:17, מאת Jarek Potiuk <ja...@potiuk.com>: > Maybe just `NukeOperator` simply :) > > On Tue, Mar 28, 2023 at 7:57 AM Daniel Standish > <daniel.stand...@astronomer.io.invalid> wrote: > > > > Happy to see the engagement on this one. Thanks to everyone for thinking > > it through and contributing their thoughts. > > > > re niko > > > > > > > - Context managers: > > > I found most of the context manager syntax proposals a little hard to > > > understand, but some better than others. Ultimately if I put my DAG > author > > > hat on, I find this declaration the most straightforward, clear and > it's > > > easy to update existing code: > chain(create_notification_channel.as_setup(), > > > ... other tasks ... > > > > delete_notification_channel.teardown_for(create_notification_channel),...) > > > > > > +1 yeah i like this approach and, exactly -- it's the simplest syntax for > > updating existing code to use the feature and requires essentially no > > change in dag structure. > > > > re short circuit > > > > 1. it is a classic 80s movie 2. yeah no objections let's make short > circuit > > not skip teardowns, it seems like a reasonable default... that said > > were someone down the line to create a PR with an option that let's you > do > > it, in a manner that is very clear about the consequences, or maybe > instead > > a distinct ShortCircuitNuclearOperator, it would not bother me. and if > > either of those were to be rejected, it would not be hard to implement > such > > for your own project, or achieve same effect through other means. > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@airflow.apache.org > For additional commands, e-mail: dev-h...@airflow.apache.org > >
