feat(mpz-garble): pre-commit inputs #149
Conversation
| OTS: OTSendEncoding<Ctx> + Send, | ||
| OTR: OTReceiveEncoding<Ctx> + Send, | ||
| { | ||
| let assigned = self.state().memory.drain_assigned(inputs); |
There was a problem hiding this comment.
Seems like a good idea to check here and return an error if the inputs were not assigned prior to calling commit
and also for the 2 methods below
There was a problem hiding this comment.
Based on our DMs, let's use the "commit if possible" approach. Then the check is not required. Let's add a clarifying comment here and in the 2 methods below like e.g.
///
/// If some of the values in `inputs` were committed earlier, this method will ignore them.
| Ok(()) | ||
| } | ||
|
|
||
| /// Commits the provided values for verifying. |
There was a problem hiding this comment.
Reading the comment may seem confusing since the verifier is not actually committing to anything, instead maybe
| /// Commits the provided values for verifying. | |
| /// Enables the prover to commit the provided values for verifying. |
There was a problem hiding this comment.
Think of it more like a "database commit" instead of cryptographic commitment. The two parties have a shared virtual machine and are jointly committing values to its memory.
There was a problem hiding this comment.
Ideally, we should avoid leaking the VM impl details into the DEAP protocol API, if possible.
The person who follows the DEAP spec doc will be confused by the sudden overload of the term "commit", imo. To avoid the confusion, we need to explain the term "commit" from the cryptographic commitment pov.
sinui0
force-pushed
the
feat/mpz-garble-commit
branch
from
0d39556
to
2a60cec
Compare
This PR exposes the ability to commit assigned values in memory.