Update release-cycle.json URL

[hugovk]

This file is now generated and published from the PEPs repo (python/peps#4331) and will be removed from the devguide soon (python/devguide#1685).

[hugovk]

PS you can add support for 3.14 and drop EOL 3.8 and 3.9 :)

[RafaelWO]

Awesome, thanks @hugovk! Also for the typo-fix 😆

[RafaelWO]

PS you can add support for 3.14 and drop EOL 3.8 and 3.9 :)

Will do, thanks for reminding!

---

May I ask a related question: do you know if future EOL dates of Python versions, e.g. 3.10, 3.11, will always go EOL on the end of October? Because I noticed that for non-yet EOLed version, the end_of_life has no day specified yet, for example 3.10 -> 2026-10.
But already EOLed versions have a day (obviously), like for 3.9 it is 2025-10-31 and for 3.8 it is 2024-10-07 (not end of month).

I just ask because I am unsure whether I should use the end or start (or even the middle?) of the future EOL month to calculate the days until EOL.

[hugovk]

Thanks!

---

The actual EOL depends on when the release manager wants to declare it EOL.

There was a 3.9 on the 9th, but it wasn't declared the final because there was a chance another release would happen, and in the end there was one on the last day of October:

• https://discuss.python.org/t/the-final-python-3-9-security-fix-release-is-out/104666

But's 3.8 final was on the 7th:

• https://discuss.python.org/t/python-3-13-0rc2-3-12-6-3-11-10-3-10-15-3-9-20-and-3-8-20-are-now-available/63161

It wasn't known for sure that would be the final 3.8:

Python 3.8 is very close to End of Life (see the Release Schedule). Will this be the last release of 3.8 ever? We’ll see… but now I think I jinxed it.

And theoretically it could be possible that a security release was made in September and there wasn't one needed in October, or I guess a final release in early November release is possible, if something really exceptional happened, or the RM wasn't available at the end of October or something.

It's why we write "Dates shown in italic are scheduled and can be adjusted" at https://devguide.python.org/versions/

But from a practical perspective, I think it's reasonable to put the last of October, as is done at https://endoflife.date/python

[RafaelWO]

Thanks for the thorough response, @hugovk!

And theoretically it could be possible that a security release was made in September and there wasn't one needed in October, or I guess a final release in early November release is possible, if something really exceptional happened, or the RM wasn't available at the end of October or something.

Yeah, it is totally understandable that you cannot fix the day in advance.

But from a practical perspective, I think it's reasonable to put the last of October, as is done at https://endoflife.date/python

Great suggestion, I will adjust Pirel accordingly 🙂