Stop pulling unused cffi as a direct dependency in core packaging metadata
Fixes #86
@art049 the CI needs approval plz
- I didn't clone the repo locally so no uv lock updated. If somebody gets to it before I do — feel free to push a commit into this branch.
Seems there's a bit of the problem in the uv setup.
error: The lockfile at `uv.lock` needs to be updated, but `--locked` was provided. To update the lockfile, run `uv lock`.
Perhaps cffi needs to be removed from the uv.lock file?
Yes
@webknjaz if you don't have the time currently I would be willing to fork your patch-1 and then commit it. I can't seem to create a reviewer's suggestion. But I could quickly fork and pull request to your branch.
Sounds good, thanks!
@art049 could you run the CI again?
@art049 do you happen to know how to remove the cffi dependency by any chance?
Caused by: Dependency `cffi` has missing `version` field but has more than one matching package
even I'm having trouble making edits and I use uv a lot for many of my numerous libraries.
@art049 do you happen to know how to remove the cffi dependency by any chance?
Caused by: Dependency `cffi` has missing `version` field but has more than one matching packageeven I'm having trouble making edits and I use uv a lot for many of my numerous libraries.
I think it's caused by an old version of uv:
Successfully installed uv version 0.5.20
I opened https://github.com/CodSpeedHQ/pytest-codspeed/pull/95.
INTERNALERROR> RuntimeError: Failed to load instrument hooks library: No module named '_cffi_backend'
@Vizonex looks like the dep should be added to the test deps. Would you like to give it a try?
Hey @webknjaz, @Vizonex, @edgarrmondragon. After taking a closer look it was not that easy to remove cffi and keep the library working.
I opened #96 with a WIP for switching to a native extension. It still needs some thorough testing, but otherwise it should be good.
@adriencaccia Thanks. This will definitely save both me and webknjaz a headache.