Bandersnatch tweaks after backend update #1482
Conversation
| return false | ||
| }; | ||
| let preouts: [bandersnatch_vrfs::VrfPreOut; N] = | ||
| core::array::from_fn(|i| self.outputs[i].0.clone()); |
There was a problem hiding this comment.
How big is the thing being cloned here? I tried to follow the types and manually see how big it is, but there are so many layers of abstraction between what's used here and the types which actually store the raw bytes (and some broken docs.rs builds in between :P) that I just gave up. (As far as I can see this has some inline big ints inside, so I'd like to know how big exactly.) Can you core::mem::size_of in a throwaway #[test] and paste me the result?
There was a problem hiding this comment.
Each element is a Bandersnatch curve group element (aka curve point).
struct Affine {
x: U256, // 32 bytes
y: U256, // 32 bytes
infinity: bool,
}
The in memory size for a 64 bit machine is 72 bytes (8 bytes are given to the infinity bool for alignment).
The constructed array size is bounded by the max number we support at the moment (MAX_VRF_IOS = 3).
There was a problem hiding this comment.
@burdges can we implement Copy for VrfInput and VrfPreOut (as the inner types are Copy)?
There was a problem hiding this comment.
Alright sure. Arkworks & dalek both have copy types of this size.
In memory, yeah affine curve points are two base field elements plus a flag. This one has a 32 byte base field, and serializes in 33 bytes.
Yes, arkworks has layers itself. And I made it worse for other reasons.
| .expect("ring-proof serialization can't fail"); | ||
| .proof | ||
| .serialize_compressed(signature.signature.as_mut_slice()) | ||
| .expect("ring-signature serialization can't fail"); |
There was a problem hiding this comment.
There is the old rule of "proof or remove" ;)
There was a problem hiding this comment.
I added some more assertions in the "backend sanity checks" test.
This test basically checks our assumptions about the serialized lengths of some objects.
These serialized lengths are constant and any variation of the sizes is early catched by the test
|
|
||
| thin_signature | ||
| .proof | ||
| .serialize_compressed(signature.signature.0.as_mut_slice()) | ||
| .expect("serialization can't fail"); |
There was a problem hiding this comment.
ArkScale requires you "audit" arkworks types for this, but doing so is simply a grep command:
https://github.com/w3f/ark-scale/blob/master/src/lib.rs#L126
https://github.com/w3f/ark-scale/blob/master/README.md
Also VrfSignature became a scale type:
https://github.com/w3f/ring-vrf/blob/master/dleq_vrf/src/traits.rs#L97
There was a problem hiding this comment.
Oh you mean the length, yeah a test for this makes sense.
* sync Westmint to Millau * "Westend parachains at Millau" dashboard
* sync Westmint to Millau * "Westend parachains at Millau" dashboard
* sync Westmint to Millau * "Westend parachains at Millau" dashboard
* sync Westmint to Millau * "Westend parachains at Millau" dashboard
* sync Westmint to Millau * "Westend parachains at Millau" dashboard
* sync Westmint to Millau * "Westend parachains at Millau" dashboard
* sync Westmint to Millau * "Westend parachains at Millau" dashboard
* sync Westmint to Millau * "Westend parachains at Millau" dashboard
* sync Westmint to Millau * "Westend parachains at Millau" dashboard
* sync Westmint to Millau * "Westend parachains at Millau" dashboard
* sync Westmint to Millau * "Westend parachains at Millau" dashboard
* sync Westmint to Millau * "Westend parachains at Millau" dashboard
* sync Westmint to Millau * "Westend parachains at Millau" dashboard
As per title