ByronBlock / MultiEraBlock explicit era tag

[rooooooooob]

trafficstars

In Byron the era tag wrapping struct is explicitly a part of the CDDL:

byron_block = [
     ; @name ebb_wrapper
     tag: 0, block: byron_eb_block //
     ; @name main_block_wrapper
     tag: 1, block: byron_main_block
]

For all the other eras there is no mention at all of this outer wrapping struct e.g. Conway (all other eras are similar):

block =
  [ header
  , transaction_bodies         : [* transaction_body]
  , transaction_witness_sets   : [* transaction_witness_set]
  , auxiliary_data_set         : {* transaction_index => auxiliary_data }
  , invalid_transactions       : [* transaction_index ]
  ]

When we retrieve blocks from oura/cardano-net-fetcher we get even these blocks with explicit era tags as if it were:

conway_block = [7, block]

One approach in order to respect the CDDL it could make sense to keep the tag info in the byron blocks but not touch them in the other eras, and then in MultiEraBlock do the check as if it were defined like:

multi_era_block =
    byron_block            ; @name byron
  / [2, shelley_block]     ; @name Shelley
  / [3, allegra_block]     ; @name Allegra
  / [4, mary_block]        ; @name Mary
  / [5, alonzo_block]      ; @name Alonzo
  / [6, babbage_block]     ; @name Babbage
  / [7, block]             ; @name Conway

or similar.

We currently have MultiEraBlock::from_explicit_network_cbor_bytes() to parse using this era tag struct.

[SebastienGllmt]

When we retrieve blocks from oura/cardano-net-fetcher we get even these blocks with explicit era tags as if it were:

As a reminder, the Cardano network specification has a separate CDDL definition from the ledger specification and CML currently doesn't contain any codegen from the networking spec. I'm not sure the networking cddl is a particularly good state (I've never heard good things about it)