You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
but also track support of Python version in mainline Linux distros. See below.
and also consider the context of a given project (library vs. application). Libraries may need a wider support of versions, while apps can have narrower ranges.
and also consider which versions in which packaging are getting major bug and security fixes.
What this mean is that the Python.org EOL'ed versions may still need to supported.
Alternatively if we start promoting only containerized deployments, we could more readily deprecated and drop older Python versions, including non EOL'ed versions, but to the detriment of using the code a libraries....
We should also consider the many small things related to change of supported versions:
CHANGELOG entries, READMEs and documentation updates
pyproject/setup changes to add version constraints?
Ci test suites across versions
One approach would be to have two policies:
for libraries, we could support all Python versions that are receiving security and bug fixes, as of today, 3.9 to 3.12
for apps, we could support only container deployments with a more restricted number of more recent versions, like the latest stable only 3.12
for new versions we should define when we adopt them as supported. For instance 3.13 has just been released, and we could state that we work to support a new major release (like 3.13) within 12 months of its publication. Or we could also support re-releases, but frankly this is not something I look forward to.
It would be good to have a simple plan to deprecate old Python versions and support new Python versions that would:
What this mean is that the Python.org EOL'ed versions may still need to supported.
Alternatively if we start promoting only containerized deployments, we could more readily deprecated and drop older Python versions, including non EOL'ed versions, but to the detriment of using the code a libraries....
We should also consider the many small things related to change of supported versions:
One approach would be to have two policies:
As for supported versions, here is the https://devguide.python.org/versions/ EOL timeline:
And here is the distro support for various versions:
The text was updated successfully, but these errors were encountered: