Smart Contract Testing

Smart Contract Testing

When to use this skill

Use when the user asks about:

• Writing tests for smart contracts
• Setting up a testing framework (Foundry or Hardhat)
• Fuzz testing or invariant testing strategies
• Fork testing against mainnet state
• Gas benchmarking and optimization verification
• Debugging failing contract tests

Testing Framework Selection

Foundry vs Hardhat

Recommendation: Use Foundry for new projects. Use Hardhat when you need complex JS scripting or specific Hardhat plugins.

Testing with Foundry

1. Unit Test Structure

Follow this pattern for each test:

// SPDX-License-Identifier: MIT
pragma solidity ^0.8.24;

import {Test} from "forge-std/Test.sol";
import {MyContract} from "../src/MyContract.sol";

contract MyContractTest is Test {
    MyContract internal target;
    address internal alice = makeAddr("alice");
    address internal bob = makeAddr("bob");

    function setUp() public {
        target = new MyContract();
        deal(alice, 100 ether);
    }

    function test_deposit_succeeds() public {
        vm.prank(alice);
        target.deposit{value: 1 ether}();
        assertEq(target.balanceOf(alice), 1 ether);
    }

    function test_deposit_revertsWhenZero() public {
        vm.prank(alice);
        vm.expectRevert(MyContract.ZeroAmount.selector);
        target.deposit{value: 0}();
    }
}

Naming conventions:

• test_functionName_scenario for expected-pass tests
• test_functionName_revertsWhen_condition for expected-revert tests
• testFuzz_functionName for fuzz tests

2. Fuzz Testing

Let Foundry generate random inputs to find edge cases:

function testFuzz_deposit(uint256 amount) public {
    // Bound the input to realistic ranges
    amount = bound(amount, 1, 100 ether);
    
    deal(alice, amount);
    vm.prank(alice);
    target.deposit{value: amount}();
    
    assertEq(target.balanceOf(alice), amount);
}

Best practices for fuzz tests:

• Use bound() to constrain inputs to valid ranges
• Use vm.assume() sparingly — prefer bound() to avoid too many rejected runs
• Set meaningful fuzz runs in foundry.toml: [fuzz] runs = 1000
• Focus fuzz testing on functions with numerical inputs and complex math

3. Invariant Testing

Define properties that must always hold regardless of actions:

contract MyContractInvariant is Test {
    MyContract internal target;
    Handler internal handler;

    function setUp() public {
        target = new MyContract();
        handler = new Handler(target);
        targetContract(address(handler));
    }

    function invariant_totalSupplyMatchesBalances() public view {
        assertEq(
            target.totalSupply(),
            target.balanceOf(address(handler))
        );
    }
}

Common invariants to test:

• Total supply equals sum of all balances
• Contract balance always covers total deposits
• Exchange rates are monotonically non-decreasing
• Access control roles cannot be self-granted

4. Fork Testing

Test against real mainnet state:

function test_swapOnUniswap_fork() public {
    // Fork mainnet at a specific block
    vm.createSelectFork("mainnet", 19_000_000);
    
    address usdc = 0xA0b86991c6218b36c1d19D4a2e9Eb0cE3606eB48;
    address weth = 0xC02aaA39b223FE8D0A0e5C4F27eAD9083C756Cc2;
    
    deal(usdc, alice, 10_000e6);
    
    vm.startPrank(alice);
    // ... execute swap logic
    vm.stopPrank();
    
    assertGt(IERC20(weth).balanceOf(alice), 0);
}

Fork testing tips:

• Pin to a specific block number for reproducibility
• Use deal() to set up token balances without needing to acquire them
• Test against multiple chains by configuring RPC URLs in foundry.toml
• Be mindful of rate limits on RPC providers

5. Gas Benchmarking

Track gas usage to catch regressions:

# Generate gas snapshot
forge snapshot

# Compare against previous snapshot
forge snapshot --check

In tests, measure specific operations:

function test_deposit_gasUsage() public {
    vm.prank(alice);
    uint256 gasBefore = gasleft();
    target.deposit{value: 1 ether}();
    uint256 gasUsed = gasBefore - gasleft();
    
    // Assert gas is within expected range
    assertLt(gasUsed, 50_000, "deposit uses too much gas");
}

6. Coverage and CI

Ensure comprehensive test coverage:

• Run forge coverage to identify uncovered code paths
• Target 100% line coverage on security-critical functions
• Set up CI to run tests on every PR
• Include gas snapshot comparison in CI to catch regressions
• Run fuzz tests with higher iterations in CI (10,000+) vs local development (256)

7. Debugging Techniques

When tests fail:

• Use forge test -vvvv for full execution traces
• Add console.log() via import "forge-std/console.sol";
• Use forge debug for step-by-step EVM execution
• Check event emissions with vm.expectEmit()
• Verify storage values with vm.load(address, slot)