fix indentation of code blocks, keeping them at 69 columns

[conradoplg]

RFC must have at most 72 columns. Since code blocks are indented with 3 spaces, this means that they must have at most 69 columns.

The text parts (inputs, outputs) were indented manually. The Python parts were indented using autopep8. Some parts may look weird but that's how they are usually indented in Python. Let me know if anything is too strange, it's likely possible to change it by e.g. adding more parenthesis.

To see that everything have been fixed, just run make and see that there are no warnings are output.

Note that it was also giving a warning Warning: Expected a valid submissionType (stream) setting, one of IETF, IAB, IRTF, independent, but found None. Will use 'IETF'. Thus I explicitly set it to IETF.

Closes #406

[chris-wood]

@conradoplg I don't feel strongly about this being valid Python, but if that's the motivation, can we make the reference implementation match this exactly?

[conradoplg]

@conradoplg I don't feel strongly about this being valid Python, but if that's the motivation, can we make the reference implementation match this exactly?

I can, but that wouldn't accomplish much, because they already differ, e.g. spec:

def compute_binding_factors(commitment_list, msg):
  msg_hash = H4(msg)
  encoded_commitment_hash = H5(
      encode_group_commitment_list(commitment_list))
  rho_input_prefix = msg_hash || encoded_commitment_hash

  binding_factor_list = []
  for (identifier, hiding_nonce_commitment,
       binding_nonce_commitment) in commitment_list:
    rho_input = rho_input_prefix || G.SerializeScalar(identifier)
    binding_factor = H1(rho_input)
    binding_factor_list.append((identifier, binding_factor))
  return binding_factor_list

While the reformatted reference implementation would look like this:

def compute_binding_factors(G, H, commitment_list, msg):
  msg_hash = H.H4(msg)
  encoded_commitment_hash = H.H5(
      encode_group_commitment_list(
          G, commitment_list))
  rho_input_prefix = msg_hash + encoded_commitment_hash

  binding_factors = {}
  rho_inputs = {}
  for _, (i, D, E) in enumerate(commitment_list):
    rho_input = rho_input_prefix + G.serialize_scalar(i)
    binding_factor = H.H1(rho_input)
    rho_inputs[i] = rho_input
    binding_factors[i] = binding_factor
  return binding_factors, rho_inputs

Also it seems better to keep using 4-space identation in the reference code for readability.

[chris-wood]

Yeah, they differ now, but I thought it might be nice if they were identical. That way there's no question about inconsistencies.

[conradoplg]

Yeah, they differ now, but I thought it might be nice if they were identical. That way there's no question about inconsistencies.

What I meant is that they are different in content, not only in formatting, e.g. H.H5 vs H5, one uses dict and the other uses list, one receives G and H and the other not, etc. I'm not sure we can make then fully equal. We can try making them as similar as possible, but that seems it would work better in a separate PR

[chris-wood]

Ah, yes, right! I'm going to mull on this for a bit, if that's OK? There's no rush to merge it.

[chris-wood]

OK, after thinking about this, I'm totally comfortable trying to address parity in a separate PR, keeping this one focused on the original width issue. I can own the followup issue on parity.