Nathan Goldbaum

@alex it looks like all the wheel builds are successful except for win32 and armv7. See these actions runs on my fork: win32: https://github.com/ngoldbaum/cryptography/actions/runs/16601346439/job/46961605807?pr=2 armv7: https://github.com/ngoldbaum/cryptography/actions/runs/16601346439/job/46961606173?pr=2 Not sure what to...

> it occurs to me that the gil_used=false should be good to split off from this and land now I'll put this in along with the new CFFI version constraints...

> At this point I think we're just waiting for the cffi release! Thanks for your work here! While the remaining build issues here aren't very fun I've been telling...

OK win32 is fixed. That came down to uv not selecting free-threaded python unless you explicitly opt into it. I added `--python` where appropriate. I tried building the armv7 wheel...

I reported a bug against maturin: https://github.com/PyO3/maturin/issues/2701

> is there any chance it's an issue in the manylinux image, and that the 3.14-freethreaded binary in the armv7l image is accidentally aarch64? looks like that's indeed what's happening:...

Apologies for the noise here, I'm trying to debug the armv7 issue.

Looks like that fixed it! There's one unrelated failure. I'll go ahead and do a separate PR for the `manylinux-entrypoint` fix.

OK, I think this should be ready whenever CFFI does a release. I'll try to rebase every now and then since this will bitrot quickly whenever the CI or wheel...

Tornado builds fine on the free-threaded build. I think @lysnikolaou looked at this a little already and has some thoughts but I'll let him clarify.