Closed tjade273 closed 8 months ago
Ugh, well here I thought I had everything working with some consistent rules for sizes, but 3293 and 4595 definitely violate that. It's a bit frustrating given hybrid-array
's tradeoffs versus generic-array
where we do need to write ArraySize
impls for each size.
That said, I guess we can just take one off sizes to support by request.
The FIPS SLH-DSA and ML-DSA draft standards have signature sizes which are very large in the former and odd sizes in the latter, so neither has signature sizes covered by the current ArraySize implementations
SLH-DSA has signatures of lengths:
ML-DSA has signatures of lengths:
ML-DSA also has private keys of size 4864 for category 5.
I don't know if you want to set a precedent for people requesting random extra ArraySize impls, but these would be useful for writing clean no-std implementations of the new PQ schemes.