dense_mlpoly.rs: Fix manipulation of evaluation vector Z by bound functions

[asn-d6]

The bound functions were folding the Z vector on itself but were not actually truncating it (even though they were changing self.len).

Hence, if you called a bound function and then evaluate() it would assert because evaluate() checks the size of the Z vector.

[vercel]

@asn-d6 is attempting to deploy a commit to the Nexus Labs Team on Vercel.

A member of the Team first needs to authorize it.

[sjudson]

@asn-d6 thanks for the PR. We'll take a look internally, but I should note as well that this Spartan code is drawn from MSR's reference implementation so you might want to also send this PR upstream there.

EDIT: There also appears to be a failing check in the CI.

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
nexus-docs	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Sep 24, 2024 1:21pm

[asn-d6]

@asn-d6 thanks for the PR. We'll take a look internally, but I should note as well that this Spartan code is drawn from MSR's reference implementation so you might want to also send this PR upstream there.

EDIT: There also appears to be a failing check in the CI.

Good point. Opened a PR upstream here: microsoft/Spartan#73

Also pushed a commit that fixes the CI, I think.

[sjudson]

Tentatively, as this appears to be a completeness issue rather than a soundness or zero-knowledge issue, I'm going to hold on it for the upstream PR (which is to an actively maintained repo) to be resolved, as we'd like to retain general parity with the reference implementation.

[asn-d6]

Upstream PR got merged for what it's worth!

[sjudson]

Approved due to upstream merge.