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
Python 3.8 has fallen out of support 2 days ago, which means it's not getting anymore security updates.
This also has consequences for dependency management since an increasing number of dependencies have moved away from it, and we're not able to get the new versions anymore.
Consider making this change at the same time as #341, and after #302 has been implemented.
IMPORTANT: this is a source breaking change for 3.8 users (they'll need to upgrade to a newer version of python), make sure to include a bang in your message commit (e.g. feat!:) so a major version bump is made.
We do not need to major bump kiota itself since the generator will emit for the same API surface and we're within our support policy.
After the PR for this is merged but BEFORE the changes are released we also need to go back to any closed (not merged) dependabot PR and ask the bot to recreate the pull requests (@dependabot recreate) so we upgrade to the latest version of the dependencies.
to add data points for this one: it seems that libraries are perfectly fine with bumping the required minor version of python in a patch/minor release.
Here are some examples:
Looking at the Azure SDK guidelines, they are just providing a (outdated?) target version. No upgrade guidance. I have asked a question in the internal channels to see if any Python architect tells us following what the ecosystem is doing here is a bad idea for some reason.
I think that with all the evidence we've collected, it'd be ok to remove support for 3.8 in a minor bump. What do you think Andrew Omondi (@andrueastman)Philip Gichuhi (@Ndiritu) ? If we're in agreement, we can probably proceed with the work.
In addition to the type hint improvement, we should also look at all the ignored dependabot PRs and make sure we update all dependencies before release the version that drops support for 3.8
I think that with all the evidence we've collected, it'd be ok to remove support for 3.8 in a minor bump. What do you think Andrew Omondi (@andrueastman)Philip Gichuhi (@Ndiritu) ? If we're in agreement, we can probably proceed with the work.
Agree. We can drop the version with a minor bump given the evidence collected.
On investigating, it looks like we are also emitting the deprecated type hints in the generated code, so I've also authored a generator PR at microsoft/kiota#5974.
My suggestion for review/release order here would be to
Activity
baywet commentedon Oct 9, 2024
additionally this documentation needs to be updated https://learn.microsoft.com/en-us/openapi/kiota/quickstarts/python#required-tools
baywet commentedon Oct 9, 2024
And this one https://learn.microsoft.com/en-us/openapi/kiota/tutorials/python-azure?tabs=portal
baywet commentedon Dec 4, 2024
to add data points for this one: it seems that libraries are perfectly fine with bumping the required minor version of python in a patch/minor release.
Here are some examples:
Which means this is most likely was the eco-system expects, and it should be fine to follow that practice on our end.
andrueastman commentedon Dec 5, 2024
Seeing the same in changelogs such as
baywet commentedon Dec 5, 2024
Looking at the Azure SDK guidelines, they are just providing a (outdated?) target version. No upgrade guidance. I have asked a question in the internal channels to see if any Python architect tells us following what the ecosystem is doing here is a bad idea for some reason.
Ndiritu commentedon Dec 12, 2024
We could also consider dropping the deprecated type hint aliases in favor of the recommended types for Python 3.9+
baywet commentedon Dec 12, 2024
I did get a first reply, Azure does drop support 6 months after the python version EOL, see their calendar
Looking at azure identity they dropped support for 3.7 without a major bump. And their guidelines indicate they do follow semver
I think that with all the evidence we've collected, it'd be ok to remove support for 3.8 in a minor bump. What do you think Andrew Omondi (@andrueastman) Philip Gichuhi (@Ndiritu) ? If we're in agreement, we can probably proceed with the work.
In addition to the type hint improvement, we should also look at all the ignored dependabot PRs and make sure we update all dependencies before release the version that drops support for 3.8
andrueastman commentedon Jan 2, 2025
Agree. We can drop the version with a minor bump given the evidence collected.
baywet commentedon Jan 6, 2025
We should now expedite this, dependabot (and most likely the associated security features) is dropping 3.8 support in less than a month
https://github.blog/changelog/2025-01-06-closing-down-notice-dependabot-will-no-longer-support-python-version-3-8
andrueastman commentedon Jan 7, 2025
Taking a look into this.
andrueastman commentedon Jan 9, 2025
I believe #433 should be okay to review now.
On investigating, it looks like we are also emitting the deprecated type hints in the generated code, so I've also authored a generator PR at microsoft/kiota#5974.
My suggestion for review/release order here would be to