Closed jjnesbitt closed 11 months ago
@AlmightyYakob FYI: For the dependencies that didn't have major version bumps (i.e., not black), increasing the minimum required versions in pyproject.toml
was the wrong thing to do. If you actually look at Dependabot's PRs, you'll see they only update the lockfile, as there is no need to update pyproject.toml
unless a dependency has a major version bump or the code starts using a feature added in a later version of a dependency. This PR causes problems for me in a project that uses aiobotocore, which has very restrictive version requirements for botocore, a dependency of boto3.
I see. Would reverting the version bumps in pyproject.toml
(except for black
) fix the issue for you? I can bump the version (0.3.0
-> 0.3.1
unless you have any objections) and yank 0.2.12
from PyPI.
0.3.1 will be useful later (The project in question doesn't yet support Pydantic v2.). Don't yank 0.2.12; pip's dependency resolver seems to do the right thing at the moment.
Since several PRs were opened, and I'd need to bump a version for each one to merge it, I bundled them all here.