Add get_group_name to expose SSL_get0_group_name

[toddabrams]

See corresponding PR in cryptography pyca/cryptography#13384

Since this PR depends on the corresponding one in cryptography, I anticipate needing to mark this as a Backward-incompatible change, once we know the cryptography version that includes the new change.

[reaperhulk]

This will be difficult to test until a cryptography release. That said, the cryptography-main job is currently failing because pyOpenSSL unconditionally assumes the binding will be there even when cryptography is linked against an OpenSSL that does not support it (in this case, that job links against system OpenSSL 3.0.2).

[toddabrams]

@reaperhulk Thanks for reviewing the cryptography PR and your comment here!

• re: difficult to test -> I was able to work around locally by installing cryptography main branch, wasn't too bad.
• re: unconditional assumption -> I’ve added a decorator that acts as a guard / conditional to address this using a pre-existing pattern 32d979a

After the next cryptography release, I’ll make the necessary changes so that the new version of pyOpenSSL requires that version of cryptography.

[toddabrams]

@reaperhulk This is 98% ready, I just need to figure out how to require cryptography 46.0.0 beyond what I've done in 9a06dc8

[alex]

Whatever the 3.14 bug is, we're also seeing it on other PRs, so it's not your fault.

[toddabrams]

Ah, thanks for letting me know, I'll ignore that one for now. I only need to worry about the 3.7 issue ERROR: No matching distribution found for cryptography==46.0.0.

[reaperhulk]

That issue has an easy fix: we dropped python 3.7 in cryptography 46 so we'll drop python 3.7 here when we bump the bound 😄

[toddabrams]

That issue has an easy fix: we dropped python 3.7 in cryptography 46 so we'll drop python 3.7 here when we bump the bound 😄

@reaperhulk Should we drop python 3.7 here or in a different PR? Is there something else still pending?

I added commits here just in case: cb720a2 5ef8f20

[reaperhulk]

Putting it here is fine.

[toddabrams]

Are there any other steps we're waiting on before this can be merged and included in a new release?

[alex]

@reaperhulk Didn't think really about the API design, but the impl looks good. If the API works for you, plz merge.