Update CI.yaml for Py3.12 and drop Py3.8 #98
Conversation
|
@orionarcher the blocker for Python 12 here is the pandas upper pin. What was the reasoning behind that to begin with before I change it? |
|
@IAlibay tagging for reference. |
|
Added missing setuptools. |
| 'mdanalysis>=2.0.0', | ||
| 'pytest', | ||
| 'matplotlib', | ||
| 'setuptools', |
There was a problem hiding this comment.
Huh setuptools as a runtime dependency is really rare - what are you using it for? Usually it goes under build time dependencies.
P.S. if you drop versioneer you could drop setup.py completely I think
There was a problem hiding this comment.
using pkg_resources in python 3.12 has an explicit dependency on setuptools now. IIRC there is a replacement but I was too lazy to search it out.
There was a problem hiding this comment.
I always mess up the versioningit setup for some reason.
There was a problem hiding this comment.
Ah yeah - I'd drop pkg_resources, best thing is to use importlib. Here's the giant MDA example 😅 : https://github.com/MDAnalysis/mdanalysis/blob/develop/testsuite/MDAnalysisTests/datafiles.py
There was a problem hiding this comment.
re: versioningit setup - here's the recent change we did for waterdynamics, hopefully should give you all the things that need changing: https://github.com/MDAnalysis/waterdynamics/pull/28/files
…the residence time calculations.
…requirements.yaml
|
Thanks for the maintenance work here @hmacdope. The pin to 2.2 should kill two birds with one stone. Will merge when tests are passing. |
|
@orionarcher I think it might be good to just cut a release after this, brings us in line with NEP29 and also probably good to resolve pin on conda-forge feedstock also. Let me know what you think. |
|
Agreed and will do! |
Fixes #97
Status