Changes in Endorser Protocol

[HarshMultani-AyanWorks]

This is a WIP pull request, and it covers the changes in endorser protocol for issues defined in

1. [Endorser Protocol] Provide an option for the Endorser to write the ledger transaction #1031
2. [Endorser Protocol] Review Writing Non Secrets #1029

[codecov-commenter]

Codecov Report

Merging #1134 (8130e52) into main (e208697) will decrease coverage by 0.18%.
The diff coverage is 67.15%.

@@            Coverage Diff             @@
##             main    #1134      +/-   ##
==========================================
- Coverage   98.68%   98.50%   -0.19%     
==========================================
  Files         450      452       +2     
  Lines       24784    24855      +71     
==========================================
+ Hits        24459    24484      +25     
- Misses        325      371      +46     

[ianco]

Hi @HarshMultani-AyanWorks , it's hard to tell from the code, but the default behaviour should be that the Author writes the transaction to the ledger and updates the schema/cred def non-secrets record to their own wallet. There should be an option (maybe a new parameter in the create-request endpoint) to ask the Endorser to write the ledger transaction (in this scenario the Author still needs to write the non-secrets record to their wallet). It looks like this update moves all of this functionality to the Endorser? (Unless I'm not reading the code properly?) Let me know once you've updated the integration test, it's broken now so difficult to review ...

[HarshMultani-AyanWorks]

Hi @HarshMultani-AyanWorks , it's hard to tell from the code, but the default behaviour should be that the Author writes the transaction to the ledger and updates the schema/cred def non-secrets record to their own wallet. There should be an option (maybe a new parameter in the create-request endpoint) to ask the Endorser to write the ledger transaction (in this scenario the Author still needs to write the non-secrets record to their wallet). It looks like this update moves all of this functionality to the Endorser? (Unless I'm not reading the code properly?) Let me know once you've updated the integration test, it's broken now so difficult to review ...

Hi @ianco,
I have fixed the integration test.
Also updated the unit tests.

I checked and tested the code, the author and endorser, both can invoke the endpoint to write the transaction to the ledger. However, the transaction is written to the ledger from the author's agent and so the records are being stored in author's wallet.
Can you point me, if I am missing something?

[sklump]

There is an order to follow, for future reference:

• import native package[s]
• one blank line
• from native package[s] import item[s]
• one blank line
• import third party package[s]
• one blank line
• from third party package[s] import item[s]
• one blank line
• from relative package[s] import item[s]: farthest to closest, one blank line between levels
• alphabetical order within each grouping above.

So "from ......connections.models.conn_record ..." belongs at line 6.

• from farthest to closest

[HarshMultani-AyanWorks]

Hi @sklump,
Thanks for the feedback, I have fixed the order of import in the recent commit - 17c0548

[sklump]

So this would be:

import json
import logging
import uuid

from asyncio import shield
from time import time

from ....connections.models.conn_record import ConnRecord
from ....core.error import BaseError
from ....core.profile import ProfileSession
from ....indy.issuer import IndyIssuerError
from ....ledger.base import BaseLedger
from ....ledger.error import LedgerError
from ....messaging.credential_definitions.util import CRED_DEF_SENT_RECORD_TYPE
from ....messaging.schemas.util import SCHEMA_SENT_RECORD_TYPE
from ....storage.base import StorageRecord
from ....storage.error import StorageNotFoundError
from ....transport.inbound.receipt import MessageReceipt

from .messages.cancel_transaction import CancelTransaction
from .messages.endorsed_transaction_response import EndorsedTransactionResponse
from .messages.refused_transaction_response import RefusedTransactionResponse
from .messages.transaction_acknowledgement import TransactionAcknowledgement
from .messages.transaction_job_to_send import TransactionJobToSend
from .messages.transaction_request import TransactionRequest
from .messages.transaction_resend import TransactionResend
from .models.transaction_record import TransactionRecord
from .transaction_jobs import TransactionJob

It would be around this many imports that adding another one starts to risk duplication, if they're out of order.

[sklump]

flip them

[sklump]

... see? New code duplicates the existing TransactionJob import. Imports need sorting.

[sklump]

apart from gory imports, this looks good

[sklump]

Could you merge this with main? Then I'll merge it.

It might take some manual edits - main has been in significant churn.