Co-authored-by: Piotr Roslaniec <[email protected]>

…nding

Or more correctly, the problem must be we're not unblinding shares in tests. This makes sense because the new verifiability features for updates were designed to work on top of blinded shares, and these tests were written for the previous share update code

As suspected 2 commits ago

For the moment, just validating share update commitments

* For refresh: C_0 == 1,
* For recovery at root t: sum(r * C_i) == 1

Get rid of PrivateKeyShare and UpdatedPrivateKeyShare at the refresh level. Instead, we'll continue to deal with BlindedKeyShares, since they're semantically similar to transcripts at the pvss & dkg level

Let the refresh layer deal with private keys

Repurposed `PrivateKey.recover_share_from_updated_private_shares` code as the test function `combine_private_shares_at`, since it doesn't make sense to combine private key shares in production code.

HandoverTranscript, a.k.a. "The Baton", represents the message an incoming node produces to initiate a handover with an outgoing node. After the handover, the incoming node replaces the outgoing node in an existing cohort, securely obtaining a new blinded key share, but under the incoming node's private key.

This was meant to be a test-only function, but let's remove it to avoid misuse

For some reason, it was part of WASM bindings too

See issue #192