pypi package

[gaardiolor]

Hi,

I think there is no pypi package? Would be nice if this could be added.

I do see a kyber-py package, but it's not yours it seems. The associated github page points to https://github.com/jack4818/kyber-py , but it redirects to your repo (!) If this is unwanted, it could violate https://peps.python.org/pep-0541/ and you could possibly raise a dispute. Maybe the one who added this package does not have malicious intent, and is open to transfer ownership to you.

If you could add it (even under a different name), and add the correct package name to the documentation, I think it would avoid confusion.

Thanks!

[GiacomoPope]

I think the owner of that library wanted kyber-py to be made and at the time I was not happy to do this so they did it without me / my permission.

It's something low priority for me right now, but yes, it would be good for this to be sorted out properly to stop outdated code being hosted on pypi

[GiacomoPope]

This is the old PR where I asked for it to not be made: #4

[GiacomoPope]

@MarcPartensky could you remove the pypi package please

[MarcPartensky]

Hello I removed the package.

[GiacomoPope]

OK -- so now my real motivation to make a package is if this really helps something for @tomato42 with your projects. Generally i still feel against it, but maybe we could then keep this fairly basic and move interoperability from your other PR into another repo?

[tomato42]

@GiacomoPope I mean, I see this package just like I see python-ecdsa (https://github.com/tlsfuzzer/python-ecdsa): a working, correct implementation useful for experimentation and testing, a practical explanation how the algorithm is working. But not supposed to be used in production, not secure against side channel attacks, etc.

Not sure what you mean by moving the interoperability into another repo... Why would having a way to read and write standard file formats for the algorithm not appropriate for this repo? I haven't updated the PRs because of the active discussions in IETF.