Re-applying the blocker tag since the issue was closed.

Changing the MAGIC_NUMBER for a maintenance release requires discussion in python-dev and notification of community stakeholders.

IMO, fixing it in RC is probably the best thing to do, though it is pretty unfortunate.
@hugovk Your thoughts on this?

(cc @hroncok)

Oh. We will branch Fedora 43 tomorrow (which doubles the fun). We can handle this, it's just a tad tedious.

@hugovk I'd appreciate some extra time for our rebuild, ideally by releasing rc2 with this fix sooner than planned and adding rc3 to the release schedule instead. Alternatively, we can backport this ourselves, but then our next build of rc1 gains a rc2-like magic number, which might be confusing.

I've bumped the version in the test, and I've opened the discussion on python-dev list Discourse, as the test requested: https://discuss.python.org/t/heads-up-possible-pyc-magic-number-change-in-3-14-0rc2/102095

I don't think this is worth spinning up the release machinery -- that's a lot of work for multiple people. If we want to avoid confusion for Fedora users, maybe skip rc2 (reserving it for rc1 with this patch), and release rc3 instead. (But I'm not sure what the impact of skipping a RC version is.)

But, @hugovk, it would be good to merge this ASAP.

Thanks Petr for approving this PR and opening the discussion.

Let's merge this now. I'll have a think about doing an early rc2 and extra rc3, I'd prefer that over skipping an rc2.

I'll have a think about doing an early rc2 and extra rc3, I'd prefer that over skipping an rc2.

I am probably OK backporting this. No need for an actual release.

I've decided anyway to do rc2 on Thursday, and add an rc3 on 2025-09-16. It's better to get this out for testing early.

• https://discuss.python.org/t/early-3-14-0-rc2-and-extra-rc3/102151
• PEP 745: Move RC2 earlier and add RC3 peps#4544