I noticed that quite a few cases of GitHub CI fail...
# all-things-deployment
b
I noticed that quite a few cases of GitHub CI failing due to python libraries releasing new versions(for instance, great expectations 0.15.4 causes issues with
Copy code
from __future__ import annotations
, causing some methods to not work as expected. We usually don't lock or constrain the versions until the version issue occur, but I was wondering if it makes more sense to just specify the version? Wondering what's the considerations behind not locking the versions (mostly), as a layman python user.
f
Ive been advised by our devops not to lock versions so we always have the newest security updates etc. Worst. Advice. Ever! I learned to fix mayor and minor versions, while keeping the bugversion wide open. For instnace ge would be great_expectations==0.15.*
👍 1