Is there long term solution to resolve python package version dependencies issue?

Every time when a new Python version release, a lot of existing Python packages will fail to work due to version dependencies issues. Does the Python Committee have a long term solution to resolve this ?

For example, some Python packages owners set the package requirements list with the following:

XXXX<=3.13

and this break the package support when Python version upgraded from 3.13.x to 3.14.x

A new and major Python release is more like a hindrance than an encouragement to people who want to use the latest Python version

3rd party packages are maintained by 3rd parties. There is a certain level of expectation in the community that maintainers will start testing against beta releases in the hopes of having all packages be compatible with the release when it comes but this is not an achievable goal. It will take some months before all major packages, especially ones that are not pure Python, are updated to work with the new version.

1 Like

Upper version limits on the Python interpreter in package metadata are generally considered bad practice. See Requires-Python upper limits for an extensive discussion on this, but one of the key issues is the one you’re describing here.

As @brass75 said, this is mostly a matter for individual package maintainers to address, there’s not much we can do centrally beyond making it clear what we consider to be best practices, and trying to define standards that make it easier to do the right thing than the wrong thing.

13 Likes

There is one other thing that can be done, which is to minimize breakage between different versions of Python, which then in turn reduces friction for package authors updating their packages (ideally to the point where no update at all is required). This issue has come up in several recent threads.

2 Likes