test(script): Assert state changes for script tests

[benceharomi]

Summary

I focused on improving the tests, but also cleaned up and organised some of the scripts to make them easier to follow.
This PR resolves #65

Approach

Not all scripts store their deployed contract addresses in public variables, so I assumed those addresses weren’t accessible. To handle address checks, I added two functions:

1. A function to calculate the expected address of a contract using the deployer, salt, bytecode, and constructor arguments.
2. A function to check if a contract is deployed at a specific address.

For each test, I followed these steps:

1. Calculate the expected address of the contract to be deployed.
2. Check that no contract exists at this address.
3. Run the deployment script.
4. Check the same address again to confirm the contract was deployed.

Each test goes through all possible branches in the script and each test case covers the deployment of one contract. I also changed the scripts so they start by loading the required env vars, if an env var is missing, the script will fail right away, before starting deployments.

The test process for each script includes:

• Testing all the scenarios where the script should revert (e.g., missing env vars).
• Testing all possible branches of execution, one test case for each.

For the calldata computation scripts, the success tests basically replicate the implementation logic, but they’re useful to confirm everything works as expected, especially if the code is refactored in the future.

To reduce repeated code, I moved common test cases into the base file so they can be reused. This makes the test files cleaner and easier to read.

This approach helped me achieve 100% code and branch coverage for all tested scripts.

ENV issues

I ran into some strange behaviour with how Foundry handles environment variables. I found a closed GitHub issue explaining this problem. The issue is that if env vars are set in both setUp and the test cases, Foundry doesn’t reload them before every test, even though setUp runs before each test. For example, if one test clears an env var set in setUp, the next test might see it as empty. Since Foundry runs tests in non-deterministic order, this causes unexpected results.

To fix this, I created a separate function that can be used to set env vars at the start of each test case. This ensures that the right env vars are always set before the tests run.

Findings

There is a lot of repeated code in the scripts, as shown by the number of reused test cases. This duplication could be reduced by either:

• Merging similar scripts and using branching logic with params (e.g., the AccountHiding scripts, which are 99% identical to the non-hiding versions).
• Moving shared functions into a helper contract or library.

I also noticed that some deployment scripts use a hardcoded salt value of 0. If this wasn’t intentional, it might need to be addressed. (DeployEmailRecoveryModule, DeploySafeRecoveryWithAccountHiding, DeployUniversalEmailRecoveryModule)

[openzeppelin-code]

test(script): Assert state changes for script tests

Generated at commit: affc9dedff40f71927f866b456db860c2ce134d8

🚨 Report Summary

	Severity Level	Results
Contracts	Critical High Medium Low Note Total	0 0 0 4 17 21
Dependencies	Critical High Medium Low Note Total	0 0 0 0 0 0

---

For more details view the full report in OpenZeppelin Code Inspector