Closed gavindsouza closed 9 months ago
Okay, we can do that. Do you know how to handle the version of the library when support for a certain version is dropped? I believe have to update the major version. What do you think?
There isn't some standard or convention for this though, it depends on the project itself. For Frappe & ERPNext (ref: https://github.com/frappe/erpnext/wiki/Supported-Versions), we maintain support for multiple major version releases for X time which means the (system deps) environment won't ever change from the first release to it's EOL.
For PyPika, where there's only one linear branch, I don't think we need to create a major release for this. A minor one will do just fine (0.39.0
). If we were to drop support for a currently supported Python env like 3.10 - then that would've probably warranted a major release.
The Unit Test Action for Python 3.6 is failing on all PRs. This happens because support for
3.6
is dropped via the setup-python action (ref: https://github.com/actions/setup-python/issues/355).However, I would suggest we remove 3.6 and 3.7 (which have both reached end-of-life), from the testing matrix and add the later versions instead (ref: https://devguide.python.org/versions/#python-release-cycle).
https://github.com/kayak/pypika/blob/30574f997c80851f7e940ad09a63e14a98871dd3/.github/workflows/pythonpackage.yml#L16
We ought to change this to
[3.8, 3.9, 3.10. 3.11, 3.12]
instead.