Python: Support EC keygen without class-instance for cryptography

[RasmusWL]

No description provided.

[tausbn]

There's something about the use of type trackers here (for what is essentially data flow) that bothers me, but I don't see that as reason enough to block this PR.

I do think, however, that there should be a better solution available... 🤔

[tausbn]

Ah, I think I can articulate at least one thing that bugs me. If we know the origin, then surely we don't have to track the keysize as well. We should be able to read that out of the fact that origin is a use of a specific API node.

[RasmusWL]

There's something about the use of type trackers here (for what is essentially data flow) that bothers me

Agreed, I also had concerns about doing it this way back in #5075 ...

Ah, I think I can articulate at least one thing that bugs me. If we know the origin, then surely we don't have to track the keysize as well. We should be able to read that out of the fact that origin is a use of a specific API node.

I see your point here 👍 Let's do that in a different PR though