Eip4844 schema definitions

[zilm13]

PR Description

• api.schema.eip4844 structures will be added in separate PR to reduce size of this one
• please check carefully 2 fixes in Capella files 6f685e1 205e7c4

Fixed Issue(s)

Fixes #6388

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.

[ajsutton]

Looks good generally. I think we've missed some Capella fields in the BeaconState. Probably worth someone more familiar with the 4844 schemas to cast an eye over it as well.

[ajsutton]

Won't this will be missing all the new fields added in Capella? It may be worth just getting all the capella fields and then replacing the latest execution payload header field. Would probably be quite useful to have a utility method around that can do that replacement.

[zilm13]

Looks like I copied it from Capella and it's the same bug there. Looking for a way to cover it in a test

[zilm13]

I need SpecLogic for tests, postpone

[zilm13]

Oh, it's not bug, @tbenr helped me to figure out

[tbenr]

I'd prefer to introduce an override pattern like this (this is for capella). To always build on top of the previous and just apply the differences.

  private static List<SszField> getUniqueFields(final SpecConfig specConfig) {
    final SszField nextWithdrawalIndexField =
        new SszField(
            NEXT_WITHDRAWAL_INDEX,
            BeaconStateFields.NEXT_WITHDRAWAL_INDEX,
            () -> SszPrimitiveSchemas.UINT64_SCHEMA);
    final SszField nextWithdrawalValidatorIndexField =
        new SszField(
            NEXT_WITHDRAWAL_VALIDATOR_INDEX,
            BeaconStateFields.NEXT_WITHDRAWAL_VALIDATOR_INDEX,
            () -> SszPrimitiveSchemas.UINT64_SCHEMA);

    final Stream<SszField> overriddenBellatrixFields =
        BeaconStateSchemaBellatrix.getUniqueFields(specConfig).stream()
            .map(
                sszField -> {
                  if (sszField.getIndex() == LATEST_EXECUTION_PAYLOAD_HEADER_FIELD_INDEX) {
                    return new SszField(
                        LATEST_EXECUTION_PAYLOAD_HEADER_FIELD_INDEX,
                        BeaconStateFields.LATEST_EXECUTION_PAYLOAD_HEADER,
                        () ->
                            new ExecutionPayloadHeaderSchemaCapella(
                                SpecConfigCapella.required(specConfig)));
                  } else {
                    return sszField;
                  }
                });
    return Stream.concat(
            overriddenBellatrixFields,
            Stream.of(nextWithdrawalIndexField, nextWithdrawalValidatorIndexField))
        .collect(Collectors.toList());
  }

[tbenr]

otherwise we could end up being forced to always build on top of Altair and add keep adding everything from there for each additional fork

[tbenr]

And we can filter out fields when we will start deleting stuff (if we ever will)

[zilm13]

What about this? needs tests and refactoring, just to show an idea 9052835
Also weird place for this logic only in BeaconStateSchema

[tbenr]

decided to left it for a future PR

[ajsutton]

Yeah it's worth nothing that Capella didn't inherit any of the fields from Bellatrix - the execution payload is a different schema in Capella so needed to be replaced. 4844 is the first point where we're wanting to inherit fields unchanged that aren't common to all forks.

[tbenr]

LGTM