Go Tdd Baby Steps

Go TDD Baby Steps

Test-Driven Development using the smallest possible increments, with Go idioms.

Three Laws of TDD

1. Don't write production code until you have a failing test
2. Don't write more test code than is sufficient to fail (compilation failures count)
3. Don't write more production code than is sufficient to pass the currently failing test

These laws create a tight feedback loop: write a tiny test, watch it fail, write just enough code to pass.

Red-Green-Refactor

1. Red - Write a small test that fails
2. Green - Write the minimal code to make it pass
3. Refactor - Clean up while keeping tests green

Each cycle should take ~2 minutes. If longer, the step is too big.

The Revert Rule

If stuck or code is getting messy:

1. Revert to the last green state
2. Rethink the approach
3. Take a smaller step

Never debug longer than the cycle itself. Revert instead.

Baby Steps with Go Table-Driven Tests

Build up the test table incrementally — each row adds ONE behavior.

Step 1: Zero/empty case

func TestParseAmount(t *testing.T) {
	tests := []struct {
		name    string
		input   string
		want    int
		wantErr bool
	}{
		{name: "empty string", input: "", want: 0, wantErr: true},
	}
	for _, tt := range tests {
		t.Run(tt.name, func(t *testing.T) {
			got, err := ParseAmount(tt.input)
			if tt.wantErr {
				require.Error(t, err)
				return
			}
			require.NoError(t, err)
			assert.Equal(t, tt.want, got)
		})
	}
}

Production code (fake it):

func ParseAmount(s string) (int, error) {
	return 0, errors.New("empty")
}

Step 2: Single simple case

Add one row, generalize production code:

{name: "single digit", input: "5", want: 5},

func ParseAmount(s string) (int, error) {
	if s == "" {
		return 0, errors.New("empty input")
	}
	return strconv.Atoi(s)
}

Step 3: Edge cases, one at a time

{name: "negative number", input: "-3", want: -3},
{name: "leading zeros", input: "007", want: 7},
{name: "non-numeric", input: "abc", wantErr: true},

Each row forces at most one production code change.

Baby Steps with Subtests and Helpers

Use t.Run for progression and t.Helper for shared assertions:

func assertStack(t *testing.T, s *Stack, wantLen int, wantEmpty bool) {
	t.Helper()
	assert.Equal(t, wantLen, s.Len())
	assert.Equal(t, wantEmpty, s.IsEmpty())
}

func TestStack(t *testing.T) {
	t.Run("new stack is empty", func(t *testing.T) {
		s := NewStack()
		assertStack(t, s, 0, true)
	})

	t.Run("push one element", func(t *testing.T) {
		s := NewStack()
		s.Push(42)
		assertStack(t, s, 1, false)
	})

	t.Run("pop returns last pushed", func(t *testing.T) {
		s := NewStack()
		s.Push(42)
		got, err := s.Pop()
		require.NoError(t, err)
		assert.Equal(t, 42, got)
		assertStack(t, s, 0, true)
	})

	t.Run("pop empty stack returns error", func(t *testing.T) {
		s := NewStack()
		_, err := s.Pop()
		require.Error(t, err)
	})
}

Each t.Run block is a baby step — one new behavior per subtest.

Baby Steps in Production Code

Fake it till you make it

1. Return a constant to pass the first test
2. Replace constant with a variable when the second test forces it
3. Generalize only when duplication demands it

Transformation Priority Premise

Prefer simpler transformations:

1. Constant → variable
2. Unconditional → conditional (if)
3. Scalar → collection
4. Statement → recursion/iteration
5. Value → mutated value

Choose the transformation that requires the least code change.

Signals

Doing it right

• Each cycle is ~2 minutes
• Tests drive design decisions
• Code grows incrementally, never in big leaps
• Refactoring happens in green state only
• You can revert any step and lose at most 2 minutes of work

Doing it wrong

• Writing multiple tests before making them pass
• Writing production code "you'll need later"
• Skipping the refactor step
• Tests require large production code changes
• Cycles regularly exceed 5 minutes
• Reluctance to revert because "too much work to lose"

Key Principles

• Smaller steps are always available. If stuck, the step is too big.
• The tests are the design tool. Let them guide structure.
• Refactoring is not optional. Clean code emerges from the refactor step, not the green step.
• Working software at every step. After each green, the code works and is shippable.
• Speed comes from small steps. Tiny steps are faster than big ones because you spend less time debugging.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: gonzaloserrano
• Source: gonzaloserrano/gopilot
• License: MIT

Install and usage instructions live in the source repository linked above.