test: enable spec tests related to eip-7549

[nazarhussain]

Motivation

Increase spec test coverage for Electra.

Description

• Enable the spec tests related to EIP7549.
• Add SPEC_FILTER_FORK and SPEC_FILTER_FORK environmental variable which will be read in during spec test
• Fix committeeIndex check during processAttestation()

Steps to test or reproduce

• Run all tests

[wemeetagain]

@g11tech is this correct?

[nazarhussain]

The committeeBits is a bit vector so the bitLen will always remain fixed which is MAX_COMMITTEES_PER_SLOT. Hence if a chain is starting with initial validators less than MAX_COMMITTEES_PER_SLOT the attestation check will never pass.

So we actually need to get how many comittees actually participatd using getTrueBitIndexes.

[ensi321]

The if(committeeBitsLength > committeeCount) check in electra-fork branch is wrong.

The check is supposedly to be identical to this line of the spec. The reason is although committeeBits has a set length of 64, actual number of committees might be less than that. If the 60th bit is set, but there is only 40 committees, then this check needs to throw error

Here we actually want the position of the last set bit in attestationElectra.committeeBits. If lastCommitteeIndex is greater than committeeCount, then we throw error.

[ensi321]

So something like this:

    const committeeIndices = attestationElectra.committeeBits.getTrueBitIndexes();
    if (committeeIndices.length === 0) {
        throw Error("Should have at least one committee bit set");
    } else {
        const lastCommitteeIndex = committeeIndices[committeeIndices.length - 1];
        if (lastCommitteeIndex >= committeeCount) {
            throw new Error("Committee index exceeds committee count");
         }
    }

This is beyond the scope of this PR, happy to fix this in another PR. Maybe add a todo here for now. Thanks for catching it @nazarhussain @wemeetagain

[ensi321]

The if(committeeBitsLength > committeeCount) check in electra-fork branch is wrong.

The check is supposedly to be identical to this line of the spec. The reason is although committeeBits has a set length of 64, actual number of committees might be less than that. If the 60th bit is set, but there is only 40 committees, then this check needs to throw error

Here we actually want the position of the last set bit in attestationElectra.committeeBits. If lastCommitteeIndex is greater than committeeCount, then we throw error.

[ensi321]

So something like this:

    const committeeIndices = attestationElectra.committeeBits.getTrueBitIndexes();
    if (committeeIndices.length === 0) {
        throw Error("Should have at least one committee bit set");
    } else {
        const lastCommitteeIndex = committeeIndices[committeeIndices.length - 1];
        if (lastCommitteeIndex >= committeeCount) {
            throw new Error("Committee index exceeds committee count");
         }
    }

This is beyond the scope of this PR, happy to fix this in another PR. Maybe add a todo here for now. Thanks for catching it @nazarhussain @wemeetagain

[nazarhussain]

@ensi321 As this PR is enabling the spec tests for electra/oeprations/attestations so any fix to make these pass preferrably go with this PR. So I just added a commit with the fix you suggested and it's passing the spec.

[ensi321]

looks good overall minor comment

[ensi321]

Curious why we would want to specify fork that we want to skip in a env variable SPEC_FILTER_FORK. What's the use case for this? I think defaultSkipOpts is serving quite well

[nazarhussain]

Quickly run the spec tests for some forks. Whilde debugging, if we run for all forks it take too much time.

SPEC_FILTER_FORK=electra yarn test:spec:minimal

[ensi321]

lgtm

[wemeetagain]

🎉 This PR is included in v1.22.0 🎉