-
Notifications
You must be signed in to change notification settings - Fork 7
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Targeted Python versions #33
Comments
Now that numpy/scipy/matplotlib have all moved to Python 3.9+, my plan is transition jobflow in that direction too. Especially since we rely and monty and maggma which likely constrain us to Python 3.9+ in the future. In terms of type hints, in case you're not aware, you can use: from __future__ import annotations to get most of the nice features of Python 3.10 type hints. |
Thanks @ml-evs for raising the point and @utf for providing some indications about the future plans. Even though I agree that in general it would be better to be forgiving in terms of requirements, if jobflow is already planning to drop support for 3.8 it is probably not worth spending time to make jobflow-remote compatible with it. One note about the annotations. I think I have tried to use the |
Okay, this sounds reasonable to me. I will extract some of the parts of #27 that modify the CI so we can at least see robustly what is working right now and what isn't, then continue with making sure that at least 3.9+ works. |
Quick Q for @gpetretto, which Python version are we targeting with jobflow-remote? Jobflow itself is 3.8+, which is what we have been testing until recently, but there is quite a lot of 3.10+ code (just nicer type hints really). I guess given that the normal way of using jobflow-remote is inside the submission scripts, which share an environment with the real "science" code, we should try to be as forgiving as possible... If we want to stick with 3.8+ (i.e., just follow jobflow from now on) I can make a PR that "downgrades" the code.
The text was updated successfully, but these errors were encountered: