Skill: TDD Implementation
When to use this skill
-
Implementing new features
-
Major refactoring
-
Complex bug fixes
Available templates
Template Usage
templates/test_module.md
Test module structure
templates/module_structure.md
DDD structure for new module
TDD Workflow (Test-Driven Development)
Phase 1: Understand
Read existing documentation
-
GLOBAL_APP_DESCRIPTION.md for global context
-
docs/STRATEGIES.md if related to trading
Identify affected modules
-
Which DDD layer? (domain/application/infrastructure)
-
Which existing files will be impacted?
Verify architecture
domain/ → Pure business logic, no I/O application/ → Orchestration, services, use cases infrastructure/ → I/O, external APIs, persistence
Phase 2: Write tests FIRST
Fundamental rule: Tests are written BEFORE implementation code.
#[cfg(test)] mod tests { use super::*;
#[test]
fn test_new_feature_basic_case() {
// Arrange
let input = ...;
// Act
let result = new_feature(input);
// Assert
assert_eq!(result, expected);
}
#[test]
fn test_new_feature_edge_case() {
// Test edge cases
}
#[test]
fn test_new_feature_error_case() {
// Test error cases
}
}
Tests MUST fail at this stage (red phase).
Phase 3: Implement
-
Minimal code to make tests pass
-
Respect DDD architecture
-
No .unwrap() in production code
-
Use Decimal for amounts
// ❌ Avoid fn process(value: f64) -> f64 { some_operation(value).unwrap() }
// ✅ Prefer fn process(value: Decimal) -> Result<Decimal, ProcessError> { some_operation(value).map_err(ProcessError::from) }
Phase 4: Refactor
Clean the code
-
Remove duplicate code
-
Remove commented-out code and obsolete comments
-
Improve variable names
-
Simplify complex expressions
Add documentation
-
Rustdoc on public functions
-
Comments for complex logic
Verify with clippy
cargo clippy --all-targets -- -D warnings
Phase 5: Validate
Invoke the testing skill
-
cargo fmt
-
cargo clippy
-
cargo test
Invoke the documentation skill
-
Update GLOBAL_APP_DESCRIPTION.md
-
Add entry in GLOBAL_APP_DESCRIPTION_VERSIONS.md
-
Increment version
Complete example
// 1. First the test #[test] fn test_calculate_position_size() { let capital = Decimal::from(10000); let risk_pct = Decimal::from_str("0.02").unwrap(); let stop_loss = Decimal::from_str("0.05").unwrap();
let size = calculate_position_size(capital, risk_pct, stop_loss);
assert_eq!(size, Decimal::from(4000)); // 2% risk, 5% stop = 40% position
}
// 2. Then the implementation pub fn calculate_position_size( capital: Decimal, risk_percentage: Decimal, stop_loss_percentage: Decimal, ) -> Decimal { let risk_amount = capital * risk_percentage; risk_amount / stop_loss_percentage }
Anti-patterns to avoid
Anti-pattern Problem Solution
Tests after code Confirmation bias Strict TDD
.unwrap() everywhere Panics in production Error handling
f64 for money Rounding errors Decimal
Tests without assertions False positives At least one assertion per test
Modify tests to pass Hides bugs Fix the code
Commented-out code Clutters codebase Remove it (use git history)