Go Test

Go has a built-in testing framework in the testing package. It follows Go’s effective, minimalist philosophy: no magic, just code.

When to Use

• Go Projects: It is the standard. No 3rd party runner needed.
• Benchmarks: Built-in support (func BenchmarkXxx(b *testing.B)).

Quick Start

// main_test.go
package main

import "testing"

func TestAdd(t *testing.T) {
    got := Add(1, 2)
    want := 3
    if got != want {
        t.Errorf("Add(1, 2) = %d; want %d", got, want)
    }
}

Run with go test ./....

Core Concepts

Table Driven Tests

The idiomatic way to write Go tests. Define a slice of structs with input/output, then loop range over them.

tests := []struct {
    input int
    want  int
}{
    {1, 2},
    {2, 4},
}
for _, tc := range tests {
    t.Run("subtest", func(t *testing.T) { ... })
}

Subtests (t.Run)

Allows hierarchical test execution and reporting.

Helper Functions

Use t.Helper() in utility functions so that failure logs point to the test caller, not the helper line.

Best Practices (2025)

Do:

• Use testify/assert: If you hate if got != want, use the testify library for assert.Equal(t, want, got). It’s the most accepted “lib” extension.
• Run with -race: go test -race ./... to detect race conditions.
• Parallelism: Use t.Parallel() inside tests to speed up execution.

Don’t:

• Don’t use assertions for everything: Go prefers explicit error checking.
• Don’t ignore errors: If a setup step fails, use t.Fatal to stop the test immediately.

References

• Go Testing Package