Rebuild for PyPy3.8 and PyPy3.9

[regro-cf-autotick-bot]

This PR has been triggered in an effort to update pypy38.

Notes and instructions for merging this PR:

1. Please merge the PR only after the tests have passed.
2. Feel free to push to the bot's branch to update this PR if needed.

Please note that if you close this PR we presume that the feedstock has been rebuilt, so if you are going to perform the rebuild yourself don't close this PR until the your rebuild has been merged.

If this PR was opened in error or needs to be updated please add the bot-rerun label to this PR. The bot will close this PR and schedule another one. If you do not have permissions to add this label, you can use the phrase @conda-forge-admin, please rerun bot in a PR comment to have the conda-forge-admin add it for you.

_{This PR was created by the regro-cf-autotick-bot. The regro-cf-autotick-bot is a service to automatically track the dependency graph, migrate packages, and propose package version updates for conda-forge. Feel free to drop us a line if there are any issues! This PR was generated by https://github.com/regro/autotick-bot/actions/runs/2077775526, please use this URL for debugging.}

[conda-forge-linter]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[h-vetinari]

Compilation errors with pyo3... CC @mattip

error: could not compile `pyo3` due to 4 previous errors

[mattip]

The missing interfaces were added in pyO3 16.0. I don’t see an explicit pyO3 dependency in the recipe. The build is pulling in an older version. I am not sure what requires updating.

[mattip]

This is all in discussion upstream, with a pending pyO3 0.15.2 release PyO3/pyo3#2260

[h-vetinari]

Hello @alex, sorry for the direct ping.

I was wondering - if a resolution to pyca/cryptography#6924 / PyO3/pyo3#2260 is not foreseeable in the near future - would you consider pyca/cryptography#6935 to be in a state that's good enough so that we would selectively backport it here in order to be able to build cryptography (and all the packages depending on it) for pypy in conda-forge?

It would help us a bunch with unblocking the current conda-forge-wide pypy rebuild, and since we don't build for cpython 3.6 anymore, we don't care about the dropped support for that either.

Since I'm not just going to willy-nilly backport WIP patches (especially for a security-relevant package), I thought I'd ask. To be clear, I'm not asking for extra work to be done (just if the PR as is would be usable) - worst case, we'll see if the pyo3 situation can be worked out and wait for the following cryptography-release.

[alex]

@h-vetinari no problem, always happy to be ping'd. As the author of that PR I consider it ready for review, modulo the compatibility issues, but it hasn't actually gotten a review :-)

If @reaperhulk (or anyone else for that matter) reviewed that PR and it looked good to them, then I think it'd be good.

[h-vetinari]

Thanks a lot @alex!

I've made an exploratory backport, because there were some small merge conflicts, and one big one due to pyca/cryptography#6410 not having seen a release yet. I've pushed the branch with the rebase to https://github.com/h-vetinari/cryptography/tree/conda, in case someone wants to take a look, but I'd also understand if we have to wait for 37.0.

[h-vetinari]

Just in case someone is coming through and merging PRs with green CI for the migration:

Wait until pyca/cryptography#6935 is reviewed (and ideally the backport to 36.0.2 in https://github.com/h-vetinari/cryptography/tree/conda as well), or until pyo3 releases a pypy3.9-compatible version.

[reaperhulk]

If 0.15.2 happens in the next week we’re likely to also release 37.0 in that time frame. What sort of timeline are you hoping for here?

[h-vetinari]

I think a week is a reasonable wait. If it were 3-4 weeks or more I'd prefer the backport (and, if someone "signs off" on the backport, I'd still do it right away, just to unblock the 900+ child packages that are waiting on cryptography in the pypy migration).

[reaperhulk]

I can do a review of the existing PR in the next day or so.

[reaperhulk]

I've reviewed the 0.16 PR and you can consider it approved from my perspective. 😄

[h-vetinari]

I've reviewed the 0.16 PR and you can consider it approved from my perspective. 😄

Thank you very much! 😊

If I could ask one last thing, could you maybe have a look at my conflict resolution from cherry-picking pyca/cryptography@8a48132 onto tags/36.0.2? I made it explicit on this branch, i.e. cherry-pick without conflict resolution and then a separate commit to fix the conflicts (the final result used here is in https://github.com/h-vetinari/cryptography/commits/conda, as mentioned above).

It's not critical since I think the resolutions are pretty obvious (and the compilation + test suite passes here), but it would still make me feel better about the whole thing. 🙃

[reaperhulk]

The conflict fixes look fine, although obviously the actual py36 stuff in the CI yaml won't work with 0.16. I assume that doesn't matter for your use cases here though 😄

[h-vetinari]

Yeah, the CI stuff is irrelevant to us. I had originally removed it from the cherry-pick because those files weren't in the sources on PyPI, and then wanted to make sure that git diff 30f878d57fb82a26c803f3e8b00fa880e76bdd05 43506bdfdff92629027c661c89285a6eefcd4ae4 is empty.

[h-vetinari]

Thank you very much for your help on this!

[mattip]

Thanks!