Go: Table-driven unit tests

Here is the code we want to test:

package search

// Find returns the smallest index i at which x <= a[i].
// If there is no such index, it returns len(a).
// The slice must be sorted in ascending order.
func Find(a []int, x int) int {
        // binary search
        i, j := 0, len(a)
        for i < j {
                m := i + (j-i)/2
                // i ≤ m < j
                if x > a[m] {
                        i = m + 1
                } else {
                        j = m
                }
        }
        return i
}

• Put the test code in a file whose name ends with _test.go.
• Write a function TestXXX with a single argument of type *testing.T. The test framework runs each such function.
• To indicate a failed test, call a failure function such as t.Errorf.

package search

import "testing"

var tests = []struct {
        a   []int
        x   int
        exp int
}{
        {[]int{}, 1, 0},
        {[]int{1, 2, 3, 3}, 0, 0},
        {[]int{1, 2, 3, 3}, 1, 0},
        {[]int{1, 2, 3, 3}, 2, 1},
        {[]int{1, 2, 3, 3}, 3, 3}, // Here's an error.
        {[]int{1, 2, 3, 3}, 4, 4},
}

func TestFind(t *testing.T) {
        for _, e := range tests {
                res := Find(e.a, e.x)
                if res != e.exp {
                        t.Errorf("Find(%v, %d) = %d, expected %d",
                                e.a, e.x, res, e.exp)
                }
        }
}

Run the tests with go test:

$ go test
--- FAIL: TestFind (0.00s)
        search_test.go:22: Find([1 2 3 3], 3) = 2, expected 3
FAIL
exit status 1
FAIL    .../search      0.001s