GO TESTING;HOW,WHAT,WHY

Introduction

关于 Go 测试，我们应该知道测试方式（或者说测试手段）、测试对象及测试原因。

How 测试方式

测试实现

举个例子。针对字符串分割函数（如下），实现单元测试。

package goTest

import "strings"

// Split slices s into all substrings separated by sep and
// returns a slice of the substrings between those separators.
func Split(s, sep string) []string {
    var result []string
    i := strings.Index(s, sep)
    for i > -1 {
        result = append(result, s[:i])
        s = s[i+len(sep):]
        i = strings.Index(s, sep)
    }
    return append(result, s)
}

在当前目录下且一样的包名 goTest ，写一个简单的 go 测试函数，如下：

package goTest

import (
    "reflect"
    "testing"
)

func TestSplit(t *testing.T) {
    got := Split("a/b/c", "/")
    want := []string{"a", "b", "c"}
    if !reflect.DeepEqual(want, got) {
        t.Fatalf("expected: %v, got: %v", want, got)
    }
}

测试函数必须以 Test 开头， 且必须携带一个 *testing.T 参数。 t *testing.T 提供改测试函数的打印、跳过、失败功能。

测试执行

当前目录下，执行 go test ，输出如下：

> go test
PASS
ok      goTest  0.005s

如果项目中存在多个 package ，若要执行所有包的测试可以在项目根目录下使用 go test ./... ，输出如下（例子：github.com/mattn/go-sqlite3）：

> go test ./...
ok      github.com/mattn/go-sqlite3    14.693s
?       github.com/mattn/go-sqlite3/upgrade    [no test files]

代码测试覆盖率

还是以字符串分割函数为例， 获取当前代码测试覆盖率方式如下：

> go test -coverprofile=c.out
PASS
coverage: 100.0% of statements
ok      goTest    0.005s

数据显示覆盖率为 100% 。若要以 HTML 方式显示可以使用命令 go tool cover -html=c.out 。

【tip】 一行命令 cover 获取当前目录下的代码测试覆盖度。 在 ~/.bashrc 中添加如下命令：

cover () {
    local t=$(mktemp -t cover)
    go test $COVERFLAGS -coverprofile=$t $@ \
      && go tool cover -func=$t \
      && unlink $t
}

执行后获取的测试覆盖度结果如下：

> cover
PASS
coverage: 100.0% of statements
ok      goTest    0.008s
goTest/wwg_split.go:7:    Split         100.0%
total:            (statements)     100.0%

问题：测试覆盖率 100% ，结束了？

多个测试用例的情况下，使用表组测试用例装填。更改 TestSplit 如下：

func TestSplit(t *testing.T) {
    tests := []struct{
        input string
        sep string
        want []string
    }{
        {input: "a/b/c", sep: "/", want: []string{"a", "b", "c"}},
        {input: "a/b/c", sep: ",", want: []string{"a/b/c"}},
        {input: "a/b/c/", sep: "/", want: []string{"a", "b", "c"}},  // trailing sep
        {input: "abc", sep: "/", want: []string{"abc"}},
    }

    for _, tc := range tests {
        got := Split(tc.input, tc.sep)
        if !reflect.DeepEqual(tc.want, got) {
            t.Fatalf("expected: %v, got: %v", tc.want, got)
        }
    }
}

增加测试用例 trailing sep 后，执行测试，结果如下：

> go test
--- FAIL: TestSplit (0.00s)
    wwg_split_test.go:23: expected: [a b c], got: [a b c ]
FAIL
exit status 1
FAIL    goTest    0.005s

根据该结果很难一下子在表组测试用例中查出是哪条。可以将 表组测试用例实现改为 map 形式 ，具体如下：

func TestSplit(t *testing.T) {
    tests := map[string]struct{
        input string
        sep string
        want []string
    }{
        "simple": {input: "a/b/c", sep: "/", want: []string{"a", "b", "c"}},
        "wrong sep": {input: "a/b/c", sep: ",", want: []string{"a/b/c"}},
        "trailing sep": {input: "a/b/c/", sep: "/", want: []string{"a", "b", "c"}},
        "no sep": {input: "abc", sep: "/", want: []string{"abc"}},
    }

    for name, tc := range tests {
        got := Split(tc.input, tc.sep)
        if !reflect.DeepEqual(tc.want, got) {
            t.Errorf("%s expected: %v, got: %v", name, tc.want, got)
        }
    }
}

执行测试结果如下：

> go test
--- FAIL: TestSplit (0.00s)
    wwg_split_test.go:23: trailing sep expected: [a b c], got: [a b c ]
FAIL
exit status 1
FAIL    goTest    0.005s

Sub tests 使用，及 '%#v' format 使用，更改 TestSplit 如下：

func TestSplit(t *testing.T) {
    tests := map[string]struct{
        input string
        sep string
        want []string
    }{
        "simple": {input: "a/b/c", sep: "/", want: []string{"a", "b", "c"}},
        "wrong sep": {input: "a/b/c", sep: ",", want: []string{"a/b/c"}},
        "trailing sep": {input: "a/b/c/", sep: "/", want: []string{"a", "b", "c"}},
        "no sep": {input: "abc", sep: "/", want: []string{"abc"}},
    }

    for name, tc := range tests {
        t.Run(name, func(t *testing.T) {
            got := Split(tc.input, tc.sep)
            if !reflect.DeepEqual(tc.want, got) {
                t.Fatalf("%s expected: %#v, got: %#v", name, tc.want, got)
            }
        })
    }
}

测试结果如下：

> go test
--- FAIL: TestSplit (0.00s)
    --- FAIL: TestSplit/trailing_sep (0.00s)
        wwg_split_test.go:24: trailing sep expected: []string{"a", "b", "c"}, got: []string{"a", "b", "c", ""}
FAIL
exit status 1
FAIL    goTest    0.005s

更好的打印格式，可以访问：

• https://github.com/k0kubun/pp
• https://github.com/davecgh/go...
• https://github.com/google/go-cmp

使用 google/go-cmp 优化打印， 更改 TestSplit 如下：

    for name, tc := range tests {
        t.Run(name, func(t *testing.T) {
            got := Split(tc.input, tc.sep)
            diff := cmp.Diff(tc.want, got)
            if diff != "" {
                t.Fatalf(diff)
            }
        })
    }

执行测试结果如下：

> go test
--- FAIL: TestSplit (0.00s)
    --- FAIL: TestSplit/trailing_sep (0.00s)
        wwg_split_test.go:29:   []string{
                  "a",
                  "b",
                  "c",
            +     "",
              }
FAIL
exit status 1
FAIL    goTest    0.005s

修复bug后 Split 代码如下：

// Split slices s into all substrings separated by sep and
// returns a slice of the substrings between those separators.
func Split(s, sep string) []string {
    var result []string
    i := strings.Index(s, sep)
    for i > -1 {
        result = append(result, s[:i])
        s = s[i+len(sep):]
        i = strings.Index(s, sep)
    }
    **if len(s) > 0 {**
        result = append(result, s)
    }
    return result**
}

执行测试，结果如下：

> go test
PASS
ok      goTest    0.006s

> cover
PASS
coverage: 100.0% of statements
ok      goTest    0.006s
goTest/wwg_split.go:7:    Split        100.0%
total:                  (statements)    100.0%

What 测试对象

Q_1：Go 应该测试所有因子吗？

A_1：显然不是。

Q_2：何时编写测试? 1.编码完成后？ 2.编码前？ 3.其他人遍写测试，像QA、TE？ 4.项目设计人员编写测试？

A_2：编码的同时编写测试代码（TDD）Article TheThreeRulesOfTdd

Q_3：C 单元测试对象是 function ，Java 单元测试对象是 Class ，类内部的方法， Go 的单元测试对象是？

A_3：package 。测试行为，而非实施。 "The public API of a package declare this is what(行为) I do, not this is how(实施) I do it."

Why 测试原因

即使你不做代码测试，别人也会做。自己发现 issues 总比别人发现来得好，不是吗？

1. 大部分的测试(自动化)应该是开发人员自己做。
2. 手工测试不应该是你测试的主体部分，因为手工测试的复杂度为O(n)
3. 测试可以确保您始终可以运送主分支
4. 测试确定软件行为（做什么、不做什么）
5. 测试让你有信心修改他人的代码

总结

• You should write tests.
• You should write tests at the same time as you write your code.
• Each Go package is a self contained unit.
• Your tests should assert the observable behaviour of your package, not its implementation.
• You should design your packages around their behaviour, not their implementation.

【注】部分资料源于GopherChina 2019 - 'How to write testable code'