Decode a Tezos hexadecimal operation

Question

I ran tezos-baker with the -l option to see all the RPC calls it makes to construct a block. Eventually, I see this:

>>>>220: http://10.10.10.203:20001/keys/tz1efSQdjbTut8NyqE6VEj6CcyRP5mL97SNi
  "01458aa8370000975c013df5a63ff9be180cec1769d316cd8623a60b32f6d365fcf996228cc7ff1e01cc000000005da772d20487083189e81a9e5b90313b7f1da145e4bf6fdfdf0189d842652d83392169d7d500000011000000010100000008000000000000975b1c3f3042b3a504e1e5a91179a7c5e245c163cde8f94cf5407cda36e15e8dfbc60000b87319134abf41d100"
<<<<220: 200 OK
  { "signature":
      "edsigtYxuYK1FJqatLVocMo5epJZKAQdqRzR5kYqVT3A44LEuGqqo6T6EK61mdsfJ6doR2bTKypg7WZjPaofG7JLmZcwqGsvfGm" }

The payload/operation, to be signed, (in hexadecimal) begins with 01, indicating this operation is a block, followed by 458aa837, which is the chain id (babylonnet in this case). The remainder is the actual block to be signed by the http tezos-signer running on port 20001. The signature returned is shown too.

How can I decode the remaining payload to see the actual contents?

An attempt with some python code is giving unparsable results:

operationHex = payload[11:]
operationBytes = bytes.fromhex(operationHex)
print("OperationBytes:", operationBytes)

ValueError: non-hexadecimal number found in fromhex() arg at position 283

The hexadecimal number at 283 is '9'. Removing from that to the end produces a bytestring:

OperationBytes: b'\x00\tu\xc0\x13\xdfZc\xff\x9b\xe1\x80\xce\xc1v\x9d1l\xd8b:`\xb3/m6_\xcf\x99b(\xcc\x7f\xf1\xe0\x1c\xc0\x00\x00\x00\x05\xdaw- Hp\x83\x18\x9e\x81\xa9\xe5\xb9\x03\x13\xb7\xf1\xda\x14^K\xf6\xfd\xfd\xf0\x18\x9d\x84&R\xd83\x92\x16\x9d}P\x00\x00\x01\x10\x00\x00\x00\x10\x10\x00\x00\x00\x80\x00\x00\x00\x00\x00\tu\xb1\xc3\xf3\x04+:PN\x1eZ\x91\x17\x9a|^$\\\x16<\xde\x8f\x94\xcfT\x07\xcd\xa3n\x15\xe8\xdf\xbc`\x00\x0b\x871

I believe the 9-to-end is a checksum of some kind.

I'm unsure how to convert this bytestring into, what I believe should be JSON. If I hex this, I just get back what I started with, so that's no help. Attempting to base58 decode this bytestring gives error: "Invalid encoding, prefix or length mismatch" Any thoughts? Much appreciated.

Answer 1

You can use the tezos-codec binary from the latest mainnet-staging branch to parse these bytes. So, signing request are indeed prefixed by the tag 1 + <chain_id> for the block cases. The remaining part is the unsigned block header itself which can be decoded :

./tezos-codec decode 005-PsBabyM1.block_header.unsigned from 0000975c013df5a63ff9be180cec1769d316cd8623a60b32f6d365fcf996228cc7ff1e01cc000000005da772d20487083189e81a9e5b90313b7f1da145e4bf6fdfdf0189d842652d83392169d7d500000011000000010100000008000000000000975b1c3f3042b3a504e1e5a91179a7c5e245c163cde8f94cf5407cda36e15e8dfbc60000b87319134abf41d100
{ "level": 38748, "proto": 1,
  "predecessor": "BLBa1WbcNYG2qEgHAaktZTFwFyoonVX8BFeQc8E33ePK8LdZfrw",
  "timestamp": "2019-10-16T19:43:14Z", "validation_pass": 4,
  "operations_hash": "LLoaMBDyNbTVQyspAhW3ftJbwmHGS8Snq3gPpwKVxdPxM2xo34dSm",
  "fitness": [ "01", "000000000000975b" ],
  "context": "CoUrkPCSQ7MUau7vuFDq77e7nsYSeRUWiTBMmtB1NRNXwbLiFcJD",
  "priority": 0, "proof_of_work_nonce": "b87319134abf41d1" }

You may also describe its schema:

./tezos-codec describe 005-PsBabyM1.block_header.unsigned binary schema

N.B.: Your protocol might no be the same depending on which branch of the project / protocol you use. Still, you may retrieve the list of supported ids using ./tezos-codec list encodings.