x448: initial import

[baloo]

This imports https://crates.io/crates/x448 in https://github.com/RustCrypto/elliptic-curves/

[baloo]

cc @dignifiedquire @kevaundray

[baloo]

This kind of broke when we merged #1284

cc @daxpedda

I'm confused. I don't understand how https://www.rfc-editor.org/rfc/rfc7748#section-5 asks for the scalar to be a 56-bytes string and now the EdwardsScalar is 57 bytes.

Am I supposed to use the DecafScalar instead?

[daxpedda]

Apologies, I noticed indeed that this should have been DecafScalar.
I'm currently preparing another PR where I try to fix those names, because having a separate scalar for each curve when only one is special, Ed448, is a bit much.

Feel free to fix this in this PR.
Basically just replace all uses of EdwardsScalar in montgomery.rs with DecafScalar.

[baloo]

No worries, I'm not sure I'm following what you're saying, I guess I'll wait for your PR ^^

[daxpedda]

I have a new and improved PR up for MontgomeryScalar: #1306.

[tarcieri]

@baloo is there a specific reason this is still draft? It'd be good to get it landed so we make sure it gets updated when ed448-goldilocks is changed

[daxpedda]

Its currently relying on #1306.
I could try to extract MontgomeryScalar from #1306 to make it easily reviewable first if that's desirable.

[baloo]

@baloo is there a specific reason this is still draft? It'd be good to get it landed so we make sure it gets updated when ed448-goldilocks is changed

I struggled for a bit with various Scalars, #1306 fixed it for me.

I'll keep rebasing and testing along with the merges until we can get this out of draft :)

[baloo]

I guess the scalar wasn't that hard to figure out in the end.

I've sent a PR over to @daxpedda 's branch to migrate over to MontgomeryXpoint khonsulabs#1