Remove point-at-infinity for proofs/commitments

[jtraglia]

PR Description

I noticed that KZGProof and KZGCommitment were using the wrong point-at-infinity bytes. It was 48 zero bytes and it should have been 0xc0 followed by 47 zero bytes. See the G1_POINT_AT_INFINITY constant here:

https://github.com/ethereum/consensus-specs/blob/dev/specs/deneb/polynomial-commitments.md#constants

I've made this mistake several times myself. It's the uncompressed bytes which are all zero; the compressed bytes include the "infinity" flag. This only appears to have been used in test code, so no worries.

• Additionally, the two classes were providing this value differently.
  • I like the static variable style, so I switched the other one to that style.
• I believe BlobSidecar#toLogString should print the abbreviated commitment.
  • I will revert this change if this was intentional.

Documentation

• I thought about documentation and added the doc-change-required label to this PR if updates are required.

Changelog

• I thought about adding a changelog entry, and added one if I deemed necessary.

[zilm13]

Is there a way to get a test for it?

[zilm13]

This message means we don't throw in constructor if it exceeds (MAX_REQUEST_BLOCKS_DENEB * MAX_BLOBS_PER_BLOCK), we will have spec check later. So here we cap it by 2048 and later check that it's 6, so this comment is made for understanding why don't we cap here on 6 and where is actual validation.
So I'd revert this file change.

[jtraglia]

Oh, I see. I misread that. Will revert that change!

[zilm13]

this line where actual validation happens

teku/networking/eth2/src/main/java/tech/pegasys/teku/networking/eth2/rpc/beaconchain/methods/BlobSidecarsByRootMessageHandler.java

Line 86 in 94d3782

if (request.size() > maxRequestBlobSidecars) {

[jtraglia]

Is there a way to get a test for it?

Yes, I think so. I will work on that.

[zilm13]

@jtraglia it would be good to have it in spec tests. We could be not alone

[jtraglia]

@jtraglia it would be good to have it in spec tests. We could be not alone

Good idea. I'll look into that. To back that up, mev-boost had the same issue:

• Use go-eth2-client's IsInfinity() to check relay pubkey flashbots/mev-boost#630

[jtraglia]

I chatted with a coworker and we think it would be best to remove the infinity value from the two classes. It shouldn't be necessary in clients; it should only be needed internally in the KZG library. Would it be okay for me to remove these?

[zilm13]

 LGTM

[zilm13]

Merged, thank you @jtraglia