Update

Things are coming along, not 100% there yet but when done it's a script that anyone with NodeJS installed can run to get the same results. It uses some creative approaches to API queries that we'll likely use elsewhere, and just needs some final tuning and independent verification.

Per the CCIP, the criteria for someone to hit this bug is:

• the user called ccd007 between the end of the cycle and the payout transaction
• the user received citycoins back from the cycle (hits the bug / bypasses the check)
• the contract call was successful and 0 STX were returned

Analysis

To analyze this, the script takes the following steps:

• gets all ccd007 transactions from the API (configurable in code, saves to and retries from file)
• gets all ccd011 transactions from the API (same)
• gets the start/end BTC and STX block heights for each cycle (saves to and retries from file)
• gets the payout amount, height, and number of blocks for affected window from ccd011 (saves to file)
• gets the transactions that match the criteria for CCIP-016 per cycle
• prints the saved cycle data, payout data, and matching transactions
• creates a markdown table of the data for easy sharing

Examples are provided in ./examples for each file generated, and also available here

Results are stored in ./results and ignored by git.

The next step is to take those transactions and calculate what addresses were affected, and how much STX they should've received during each cycle. We have access to historical data through our Tardis contracts and can deploy something similar if needed.

To run yourself

Until this is merged you can run it by following the steps below with Node.JS installed:

1. cloning the repo (pick one)

HTTP: git clone https://github.com/citycoins/scripts
SSH: git clone [email protected]:citycions/scripts

2. checkout the feature branch

git checkout feat/ccip-016-calculation

4. install the dependencies

npm install

5. run the script

npx ts-node src/dao/ccip016-calculation.ts

The output will provide the analysis (WIP) and data files will be saved to the results folder.

Future runs of the script will read from the files to minimize API calls.

Screenshots

Loading transactions

Getting start/end block heights

Finding missed payout transactions

Example of cycle data

Example of payout data

Example of missed payout transactions
([Object] represents a transaction object)

Example of markdown output
(will be updated more)

Example of markdown output when pasted into GitHub, although the format will likely change to create a "format report" that's easy to share:

@friedger or anyone proficient with SQL, can we verify the data produced by each step? I imagine there are some queries that would achieve the same goal. Verifying the examples from each step would be a bonus.