-
Notifications
You must be signed in to change notification settings - Fork 1.6k
deprecate support for python 3.5 #5387
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
Remind me why? (No objection) |
Nobody really uses it (it's in Xenial and oldstable, but there's no real evidence of use beyond 5% downloads that are attributable to CI systems), it bloats our test matrix (which is about to expand again for Python 3.9), and dropping it gets us one step closer to good type hints (2.7 being the remaining large blocker). |
type hints was the piece I was looking for, thanks. |
Build succeeded (check pipeline).
|
Howdy! Just saw this PR go by. FWIW, we're still using Python3.5 on Xenial in SecureDrop, although we're planning a migration to Python 3.8 on Ubuntu Focal (freedomofpress/securedrop#4768) by early 2021. @reaperhulk Can you provide some more detail about the timeline for "next release" which will drop support for Python 3.5? We're extremely interested in shipping a fully patched version of cryptography, as it's a critical dependency for us. 🙂 |
The next release (3.1) will still support 3.5. We generally release every 2-3 months so that's likely to be around October. The following release (3.2) will drop support and should be around January 2021. I see in that issue you're potentially planning to go all the way up to the Xenial EOL date. Is that the real date or a worst case scenario? |
That's helpful, thanks for clarifying. You're right, we're bumping up against the EOL window pretty closely. Part of the challenge is we don't run the SecureDrop servers ourselves, so we need to work with newsrooms that are hosting them and encourage them to update. By January 2021 we plan to be quite loud about the need to upgrade, as we did with the previous LTS migration a few years back. In the event of a critical bugfix during that time window, we may well be able to backport ourselves as a stopgap solution until folks have moved over. |
If you end up in that situation please reach out to us, but it sounds like we're probably going to be in okay shape here 😄 |
This fixes a typo that's been introduced in pyca#5387.
This fixes a typo that's been introduced in #5387.
I fail to compile from source cryptography to get it working on Python-3.9.0rc1 (windows).... could it be added to the test matrix with Python-3.9rc ? and the result on pypi ? |
well, not sure where the problem is, maybe pynacl:
|
well, apparently pip installs this one on python-3.9: "cryptography-3.1-cp36-abi3-win_amd64.whl" |
No description provided.