Skip to content

Latest commit

 

History

History
95 lines (74 loc) · 2.58 KB

TXSPEC.md

File metadata and controls

95 lines (74 loc) · 2.58 KB

Transaction Specification

Format

Transactions passed to the Ledger device will be in the following format. The Ledger device MUST accept any transaction (valid as below) in this format.

{
  "account_number": {number},
  "chain_id": {string},
  "fee": {
    "amount": [{"amount": {number}, "denom": {string}}, ...],
    "gas": {number}
  },
  "memo": {string},
  "msgs": [{arbitrary}],
  "sequence": {number}
}

msgs is a list of messages, which are arbitrary JSON structures.

Examples

{
  "account_number": "123",
  "chain_id": "cosmoshub-4",
  "fee": {
    "amount": [{"amount": "4000", "denom": "uatom"}, ...],
    "gas": "40000"
  },
  "memo": "this is a comment",
  "msgs": [{arbitrary}],
  "sequence": "42"
}

Note, all the {number} values must be passed as string.

Display Logic

The Ledger device SHOULD pick a suitable display representation for the transaction.

The key type (secp256k1 / ed25519), chain_id, account_number, sequence, fee, and memo should be displayed in that order, each on their own page, autoscrolling if necessary.

msgs should be iterated through and each displayed according to the following recursive logic:

display (json, level)
  if level == 2
    show value as json-encoded string
  else
    switch typeof(json) {
      case object:
        for (key, value) in object:
          show key
          display(value, level + 1)
      case array:
        for element in array:
          display(element, level + 1)
      otherwise:
        show value as json-encoded string
    }

starting at level 0, e.g. display(msgs[0], 0).

Validation

The Ledger device MUST validate that supplied JSON is valid. Our JSON specification is a subset of RFC 7159 - invalid RFC 7159 JSON is invalid Ledger JSON, but not all valid RFC 7159 JSON is valid Ledger JSON.

We add the following two rules:

  • No spaces or newlines can be included, other than escaped in strings
  • All dictionaries must be serialized in lexicographical key order

This serves to prevent signature compatibility issues among different client libraries.

This is equivalent to the following Python snippet:

import json

def ledger_validate(json_str):
  obj = json.loads(json_str)
  canonical = json.dumps(obj, sort_keys = True, separators = (',', ':'))
  return canonical == json_str

assert ledger_validate('{"a":2,"b":3}')
assert ledger_validate('{"a ":2,"b":3}')
assert not ledger_validate('{"a":2,\n"b":3}')
assert not ledger_validate('{"b":2,"a":3}')
assert not ledger_validate('{"a" : 2 }')