Ian Stapleton Cordasco
Ian Stapleton Cordasco
This doesn't merge cleanly, doesn't solve the problem at the right layer, and it re-invents a wheel for which there are many good solutions. I'm closing this.
Do you have any intention of making this ready for review at any point?
### Thoughts on `--retry` I'm vaguely -0 on `--retry`. The only reason I'm considering it, however, is that I know that the pypi-publish-action that's maintained is something that's widely used...
> If this is a matter of tweaking the existing retry logic as @asottile-sentry suggests, I'm game to review a PR Hard to say because that's not at all what...
Actually there's a blog post from Hynek that makes lays out how to do this well. It's not hard
Yes, right now a Repository expects username/password https://github.com/pypa/twine/blob/0bb428c410b8df64c04dc881ac1db37d932f3066/twine/repository.py#L45 but we could easily refactor that, since no one has picked up on that being a publicly available API. Regardless, we can...
> Is there a reason why the response message string is checked directly instead of using the status code? Because pypi.org (and it's predecessors) as well as the other non-PyPI...
I have no clue what would cause the difference honestly. We don't send anything related to your locale in the request so even if the container was inheriting something from...
> Other indices? twine supports uploading to an arbitrary index; are there popular third-party indices that still support PGP signatures? I don't know and that's the big question I have....
Spiritually I think this is a duplicate of #430. Regardless it's part of the overall theme of "twine doesn't check things that we have to reimplement because PyPI doesn't have...