remove sgn0_be, change sgn0 to output in {0,1}

[kwantam]

This PR makes the changes we discussed in #225.

Can we hold off on merging this until we get comments from folks working to deploy BLS signatures?

Closes #225

[kwantam]

CC @JustinDrake @hoeteck @zhenfeizhang @sergeynog

For background, see #225. The goal is to significantly simplify the document by unifying the notion of "sign" for all curves. Unfortunately, this is a breaking change for BLS12-381, and we'd like feedback on how much of a killer this is.

@JustinDrake, I know I said that there were probably no more changes coming down the pipe. I'm sorry... 😖 (If it's any consolation, the change is extremely simple from an implementation point of view...)

[JustinDrake]

If it's a significant spec cleanup/simplification which is minimal for implementers then in theory it sounds good to me. Having said that I'll differ to @djrtwo who can provide much more informed feedback based on the development reality of Eth2 clients :)

cc @CarlBeek @kirk-baird @mratsim

[sergeynog]

Hi Riad: This sounds good. I don't think there will be any issues from our end. Regards, Sergey

…

On Fri, Mar 27, 2020 at 11:12 AM Justin ***@***.***> wrote: If it's a significant spec cleanup/simplification which is minimal for implementers then in theory it sounds good to me. Having said that I'll differ to @djrtwo <https://github.com/djrtwo> who can provide much more informed feedback based on the development reality of Eth2 clients :) cc @CarlBeek <https://github.com/CarlBeek> @kirk-baird <https://github.com/kirk-baird> @mratsim <https://github.com/mratsim> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#230 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGD75LVOUT6JE4KRDCWRJ7TRJS66XANCNFSM4LUX5SQQ> .

[kirk-baird]

From an implementation point of view I'm happy for removing sgn0_be and only using sgn0_le as it is simpler and faster (although the speed is negligible relative to the rest of hashing and pairing).

For Ethereum we still have not updated our specs to match hash-to-curve-06 and thus we are already introducing breaking changes so no issues on that front.

[chris-wood]

Merged after hearing no pushback!