Jarek Potiuk
Jarek Potiuk
> they may be fast enough for you but I don't consider a 100-200ms pause acceptable I know it is, you wrote it - that's why I proposed to implement...
> I would rather not persist anything to disk if possible -- it should be easy to store the computed state in shell variables (this is how the intermediate `git`...
> and re: airflow -- cool, you're free to write whatever wrappers you want especially if they save you time. admittedly airflow's usage of pre-commit is a bit non-idiomatic (hundreds...
> I don't even know why you're arguing about implementation if you have no investment in implementation or maintenance Trying to help others suffering from the same problem. I wrote...
Very much agree @flying-sheep. This would be my choice as well. I think it is quite "proper" way of doing it when you consider how much things have improved since...
Yes. you need to extend the class and add your own authentication if you want to use Kerberos.
I think maybe that's somethign that can be added in the tool we use for generation of the clients ? I woudl have found it a bit problematic if we...
Interesting approach 👀
Yep - see the comment in https://github.com/apache/airflow/issues/15635#issuecomment-1020003975, In Airflow we are getting closer to supporting M1 images (I even bought a new MacBook with this effort in mind :D). I...
Yeah. Last blocker (I believe) beam 2.36.0 has been released today for the ARM image, so :crossed_fingers: for the ARM image soon'ish