GitHub Actions: Test on Python 3.14 release candidate 2

[cclauss]

Python v3.14 -- October 7th

• https://www.python.org/download/pre-releases
• https://www.python.org/downloads/release/python-3140rc2

Note: Before submitting a code change, please review our contributing guidelines.

Description

Please describe your pull request. If it fixes a bug or resolves a feature request, be sure to link to that issue. When linking to an issue, please use refs #... in the description of the pull request.

[browniebroke]

Django 5.2 doesn't (yet) support Python 3.14: https://docs.djangoproject.com/en/dev/faq/install/

[cclauss]

https://code.djangoproject.com/ticket/35844

Django 5.2 will be the first version to support Python 3.14

The Django 5.2 tests in this PR pass on Python 3.14.

The Django main tests fail but are ignored in tox.

[browniebroke]

code.djangoproject.com/ticket/35844

Django 5.2 will be the first version to support Python 3.14

Thanks, that's what I wanted to clarify!

[peterthomassen]

lgtm!

[peterthomassen]

Once 3.14 is released, we should update the project metadata (but not before). How do we make sure we don't forget? - Or rather keep this open until then (and only use as a canary to spot failing tests if any)?

[browniebroke]

Once 3.14 is released, we should update the project metadata (but not before).

Hum... Why not before? The changes between RC2 and the final should be minimal if not none, and even if there were some changes, we'd unlikely be impacted. Let me know what I'm missing

[peterthomassen]

Well, I don't feel strongly, and I agree with you that it's unlikely. But it would be weird if we published "Python 3.14 is supported" before it's released, then the final release of it actually has an unintended (?) significant change that breaks our test, and gets released, so our stable version is broken right away. That's just not very clean, and I don't see the benefit of making the public claim early, and keeping the PR around until after doesn't seem much overhead. (Further, even if we don't update our metadata, AFAIK one could still use DRF with Python 3.14.)

But again, I don't feel strongly. We should just avoid having 2 PRs for the matrix change and the metadata change, because there's a risk to forget the second one.