Replies: 1 comment
-
+1 came here after the exact same troubleshooting steps, would love know if there's a way to prevent checking all the overrides. Even explicitely locking python/pandas/numpy versions the overrides are all still tried and resolution takes 3x longer. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I have a project where dependency resolution takes about a minute but one of the dependencies (Pandas) has a complex set of dependencies so when Poetry looks at that package it starts again from scratch and tries all the different combinations of markers so the entire process takes 5 minutes instead.
From https://github.com/pandas-dev/pandas/blob/master/setup.cfg
In my pyproject.toml I've specified a minimum version for numpy, and I thought that might stop it retrying everything. It seems to me that by telling it that it needs numpy >=1.21.0 for all platforms that should mean the platform specific resolutions are redundant:
Sadly it makes no difference setting a version for numpy, it still goes through the entire resolution process several times.
I can limit the combination by setting Python <3.10 so it only tries the first three combinations, or >=3.10 so it only tries the last one (but I have Black as a dependency also and Black has an override for Python >=3.10 so that more than doubles the total time again).
Beta Was this translation helpful? Give feedback.
All reactions