Feature: WRITE_KZG_COMMITMENT hint #176
Conversation
odesenfans
force-pushed
the
od/write-zkg-commitment
branch
2 times, most recently
from
1d93fe6 to
2d73870
Compare
odesenfans
changed the title
There was a problem hiding this comment.
In what sense we do not know what needs to be computed? Python implementation is not clear?
In case it is not already known: https://community.starknet.io/t/data-availability-with-eip4844/113065
| @@ -29,7 +30,10 @@ type StorageByAddress = HashMap<Felt252, OsSingleStarknetStorage<DictStorage, Pe | |||
|
|
|||
| /// Maintains the info for executing txns in the OS | |||
| #[derive(Debug)] | |||
| pub struct ExecutionHelper { | |||
| pub struct ExecutionHelper<KZG> | |||
There was a problem hiding this comment.
What do we gain by adding type parameter here? KZG calculation algorithm is not expected to change any time soon, neither we need multiple versions of ExecutionHelper with different KZG computers.
There was a problem hiding this comment.
Exactly, which is why a generic makes sense here. It's specified as a callback for some reason in the Python code, hinting that several implementations may be used. On the other hand, it's Rust, so it's better if we can get this indirection to be compile-time. This led me to add this as a generic. I'd be happy with having only one implementation as well but I could not find the implementation of the KZG computation anywhere.
There was a problem hiding this comment.
Over-engineering imho. Even if there are multiple implementations of kzg there is no need to have multiple ExecutionHelpers using different implementations at the same time. I would consult Ariel, maybe I am missing something.
There was a problem hiding this comment.
Python uses(in the nonpublic version): https://pydigger.com/pypi/ckzg.
It seems there is a pure rust version: https://github.com/grandinetech/rust-kzg.
Another thing to consult with Ariel.
This implementation uses a default (panicking) implementation of the KZG commitment computation as we do not know how these must be computed.
odesenfans
force-pushed
the
od/write-zkg-commitment
branch
from
2d73870 to
f3d61cd
Compare
odesenfans
changed the title
|
Superseded by #304. |
This implementation uses a default (panicking) implementation of the KZG
commitment computation as we do not know how these must be computed.
Issue Number: N/A
Type
Description
Breaking changes?