Jeff Burdges
I'm torn still because really all performant dlog crypto libraries should have keypair types that provide this optimization. It's maybe worse to hide the keypair form a person who is...
In the end, this is a performance optimization, but it's often best to stick with the usual terminology. I actually did one rename here already btw: - SecretKey -> MiniSeretKey...
Just fyi, I added a hard derivation path which previously I'd ignored.
Your derives tests look like test vectors. We've no test vectors yet because the scheme was not stable https://github.com/w3f/schnorrkel/issues/36 In this case, there were so many breaking changes between 0.1.1...
It looks like it's dose some sort of functionality tests I guess. I'm not sure about the js stuff
Interesting, we could submit a pull request to merlin, not sure if they'd accept it, but hey.. https://github.com/dalek-cryptography/merlin/pull/44 It's not afaik necessarily undefined behavior for foreign code to pass rust...
We cannot expect merlin to be changed, so we should take another route, either abusing leaking, using empty labels, or just dumping or forking merlin. I like both merlin and...
I've taken the meaningless labels approach while adding our own domain separation in https://github.com/w3f/schnorrkel/commit/fa0a6f72f46f04a8f2af83d896952a9e093d47c2 We should still benchmark merlin vs shake128 though. I also removed merlin from our bls crate...
Merged. We require 0.6.1 now since I got overzealous and yanked 0.6.0
We've an audit starting this week, so expect a breaking 0.7 before August.