Henry de Valence
Henry de Valence
There's no consensus state yet, though, because commit has never been called.
This also seems like it could be good to do for Groth16 proofs (cc @redshiftzero); I noticed in a profile that it spends a lot of time deserializing them doing...
One sub-case of this that could be worth doing sooner is the `IdentityKey`, which is currently a significant fraction of the total sync time.
We should do at least the `IdentityKey` and ZK proof changes prior to mainnet.
Should we be using metrics for this at all? I think the answer is probably no. I don't think we should be using metrics for anything that requires reconciling ongoing...
Is it desirable to use cometmock for the smoke test, though? Don't we want the smoke test to be testing the actual code we'll be running?
We can solve this problem by removing tests from the smoke test as our application-wide integration testing story matures (cc @cratelyn)
I think this subsumes penumbra-zone/penumbra#3512.
`fiat-crypto` has a web interface, neat: https://mit-plv.github.io/fiat-crypto/?argv=%5B%22word-by-word-montgomery%22%2C%22--lang%22%2C%22Rust%22%2C%22fq%22%2C%2232%22%2C%228444461749428370424248824938781546531375899335154063827935233455917409239041%22%5D&interactive
Some output for Fq: https://gist.github.com/hdevalence/6556634efe15f660919561d2d55b8443