Consider the following test:

func Test(t *testing.T) {
	for _, name := range []string{"foo", "bar", "baz"} {
		data := name // This line is necessary
		t.Run(name, func(t *testing.T) {
			t.Parallel()
			t.Log(data)
		})
	}
}

Without the name := name line to capture the name variable, the subtest will execute with the wrong values as the testdata. Worst yet, the tests will run with the correct names, but the wrong data, completely misleading the user that the test is executing as expected.

The blog on subtests explicitly warns that this must be done:

Note that we need to capture the range variable to ensure that tc gets bound to the correct instance.

I propose that vet check for this mis-use. One possible algorithm:

• If t.Parallel is called within a t.Run
• and the body of t.Run following the t.Parallel references a variable scoped to the range of some outer loop (i.e., not captured)
• then flag as potential issue.

\cc @dominikh