Implement "header matches block"

[mrBliss]

The Block Fetch logic requires the following function:

-- | Given a block header, validate the supposed corresponding block
-- body.
blockMatchesHeader :: header -> block -> Bool

We need an implementation of it for ByronHeader and ByronBlock (from Ouroboros.Consensus.Ledger.Byron).

From Slack:

Damian Nadales [1 hour ago]
updateBody checks the delegation, update, and utxo proofs

Damian Nadales [1 hour ago]
but it doesn't use the header

Edsko de Vries [1 hour ago]
But surely something somewhere must check that these match?

Damian Nadales [1 hour ago]
Oh, wait, it does use the header

Damian Nadales [1 hour ago]
here

  :: MonadError ChainValidationError m
  => BodyEnvironment
  -> BodyState
  -> ABlock ByteString
  -> m BodyState
updateBody env bs b = do
  -- Validate the block size
  blockLength b <= maxBlockSize `orThrowError` ChainValidationBlockTooLarge

  -- Validate the delegation payload signature
  proofDelegation (blockProof b)
    == hashDecoded (blockDlgPayload b)
    `orThrowError` ChainValidationDelegationProofError

Damian Nadales [1 hour ago]
where blockProof comes from the header ...

blockProof :: ABlock a -> Proof
blockProof = headerProof . blockHeader

Edsko de Vries [45 minutes ago]
so its basically

Edsko de Vries [44 minutes ago]

  proofDelegation (blockProof b)
    == hashDecoded (blockDlgPayload b)
    `orThrowError` ChainValidationDelegationProofError

Edsko de Vries [44 minutes ago]
that does the "does this thing match the header" check right?

Damian Nadales [44 minutes ago]
Yes, that'd be my understanding

Damian Nadales [44 minutes ago]
the blockDlgPayload comes from the body

Edsko de Vries [43 minutes ago]
oh

Edsko de Vries [43 minutes ago]
but that's just one part of the body

Edsko de Vries [43 minutes ago]
there are a bunch of these proofs

Damian Nadales [43 minutes ago]
yes

Damian Nadales [43 minutes ago]
I didn't paste them 😉

Damian Nadales [43 minutes ago]
but they follow a similar pattern

Edsko de Vries [43 minutes ago]
@ru do you think it would make sense to move all of these proof checks to a separate function that we could call, passing a header and a block?

Edsko de Vries [41 minutes ago]
In the consensus layer we downloaded headers and blocks separately, and then need to check "does this header indeed belong to this block"?

Damian Nadales [40 minutes ago]

@ru do you think it would make sense to move all of these proof checks to a separate function that we could call, passing a header and a block?

... and if so we might want to update the specs accordingly ....

[mrBliss]

If not done before testnet, definitely before mainnet, because it is a security risk.

[mrBliss]

See IntersectMBO/cardano-ledger#549