一个测试工程师走进一家酒吧……

原創

2021-06-25 16:28

0，t(i+1)为对于 t(i)的测试，t(i+1)正确为 t(i)正确的必要条件，那么对所有的 i，i>0，t(i)正确都是 t(0)正确的必要条件。。。"}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https:\/\/static001.geekbang.org\/infoq\/59\/59926aa33715b243e0541d2e2b60c02d.png","alt":"图片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"测试的种类"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"测试的种类有非常多，我们这里只挑几个对一般开发者来说比较重要的测试，做简略的说明。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"白盒测试、黑盒测试"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"首先是从测试方法上可以分为白盒测试和黑盒测试（当然还存在所谓的灰盒测试，这里不讨论）"}]},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"白盒测试 (White-box testing)：白盒测试又称透明盒测试、结构测试等，软件测试的主要方法之一，也称结构测试、逻辑驱动测试或基于程序本身的测试。测试应用程序的内部结构或运作，而不是测试应用程序的功能。在白盒测试时，以编程语言的角度来设计测试案例。测试者输入数据验证数据流在程序中的流动路径，并确定适当的输出，类似测试电路中的节点。"}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"黑盒测试 (Black-box testing)：黑盒测试，软件测试的主要方法之一，也可以称为功能测试、数据驱动测试或基于规格说明的测试。测试者不了解程序的内部情况，不需具备应用程序的代码、内部结构和编程语言的专门知识。只知道程序的输入、输出和系统的功能，这是从用户的角度针对软件界面、功能及外部结构进行测试，而不考虑程序内部逻辑结构。"}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"我们写的单元测试一般属于白盒测试，因为我们对测试对象的内部逻辑有着充分了解。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"单元测试、集成测试"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"从测试的维度上，又可以分为单元测试和集成测试："}]},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在计算机编程中，单元测试又称为模块测试，是针对程序模块来进行正确性检验的测试工作。程序单元是应用的最小可测试部件。在过程化编程中，一个单元就是单个程序、函数、过程等；对于面向对象编程，最小单元就是方法，包括基类、抽象类、或者派生类中的方法。"}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"整合测试又称组装测试，即对程序模块采用一次性或增值方式组装起来，对系统的接口进行正确性检验的测试工作。整合测试一般在单元测试之后、系统测试之前进行。实践表明，有时模块虽然可以单独工作，但是并不能保证组装起来也可以同时工作。"}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"单元测试可以是黑盒测试，集成测试亦可以是白盒测试"}]}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"回归测试"}]},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"回归测试是软件测试的一种，旨在检验软件原有功能在修改后是否保持完整。"}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"回归测试主要是希望维持软件的不变性，我们举一个例子来说明。例如我们发现软件在运行的过程中出现了问题，在 gitlab 上开启了一个 issue。之后我们并且定位到了问题，我们可以先写一个测试（测试的名称可以带上 issue 的 ID）来复现问题（该版本代码运行此测试结果失败）。之后我们修复问题后，再次运行测试，测试的结果应当成功。那么我们之后每次运行测试的时候，通过运行这个测试，可以保证同样的问题不会复现。"}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"一个基本的测试"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"我们先来看一个 Golang 的代码："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"\/\/ add.go\npackage add\n\nfunc Add(a, b int) int {\n return a + b\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"一个测试用例可以写成："}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"\/\/ add_test.go\npackage add\n\nimport (\n \"testing\"\n)\n\nfunc TestAdd(t *testing.T) {\n res := Add(1, 2)\n if res != 3 {\n t.Errorf(\"the result is %d instead of 3\", res)\n }\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在命令行我们使用 go test"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"go test\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"这个时候 go 会执行该目录下所有的以_test.go 为后缀中的测试，测试成功的话会有如下输出："}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"% go test\nPASS\nok code.byted.org\/ek\/demo_test\/t01_basic\/correct 0.015s\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"假设这个时候我们把 Add 函数修改成错误的实现"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":" \/\/ add.go\npackage add\n\nfunc Add(a, b int) int {\n return a - b\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"再次执行测试命令"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"% go test\n--- FAIL: TestAddWrong (0.00s)\n add_test.go:11: the result is -1 instead of 3\nFAIL\nexit status 1\nFAIL code.byted.org\/ek\/demo_test\/t01_basic\/wrong 0.006s\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"会发现测试失败。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"只执行一个测试文件"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"那么如果我们想只测试这一个文件，输入"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"go test add_test.go\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"会发现命令行输出"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"% go test add_test.go\n# command-line-arguments [command-line-arguments.test]\n.\/add_test.go:9:9: undefined: Add\nFAIL command-line-arguments [build failed]\nFAIL\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"这是因为我们没有附带测试对象的代码，修改测试后可以获得正确的输出："}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"% go test add_test.go add.go\nok command-line-arguments 0.007s\n"}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"测试的几种书写方式"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"子测试"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"通常来说我们测试某个函数和方法，可能需要测试很多不同的 case 或者边际条件，例如我们为上面的 Add 函数写两个测试，可以写成："}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":" \/\/ add_test.go\npackage add\n\nimport (\n \"testing\"\n)\n\nfunc TestAdd(t *testing.T) {\n res := Add(1, 0)\n if res != 1 {\n t.Errorf(\"the result is %d instead of 1\", res)\n }\n}\n\nfunc TestAdd2(t *testing.T) {\n res := Add(0, 1)\n if res != 1 {\n t.Errorf(\"the result is %d instead of 1\", res)\n }\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"测试的结果：（使用-v 可以获得更多输出）"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"% go test -v\n=== RUN TestAdd\n--- PASS: TestAdd (0.00s)\n=== RUN TestAdd2\n--- PASS: TestAdd2 (0.00s)\nPASS\nok code.byted.org\/ek\/demo_test\/t02_subtest\/non_subtest 0.007s\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"另一种写法是写成子测试的形式"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"\/\/ add_test.go\npackage add\n\nimport (\n \"testing\"\n)\n\nfunc TestAdd(t *testing.T) {\n t.Run(\"test1\", func(t *testing.T) {\n res := Add(1, 0)\n if res != 1 {\n t.Errorf(\"the result is %d instead of 1\", res)\n }\n })\n t.Run(\"\", func(t *testing.T) {\n res := Add(0, 1)\n if res != 1 {\n t.Errorf(\"the result is %d instead of 1\", res)\n }\n })\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"执行结果："}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"% go test -v\n=== RUN TestAdd\n=== RUN TestAdd\/test1\n=== RUN TestAdd\/#00\n--- PASS: TestAdd (0.00s)\n --- PASS: TestAdd\/test1 (0.00s)\n --- PASS: TestAdd\/#00 (0.00s)\nPASS\nok code.byted.org\/ek\/demo_test\/t02_subtest\/subtest 0.007s\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"可以看到输出中会将测试按照嵌套的结构分类，子测试的嵌套没有层数限制，如果不写测试名的话，会自动按照顺序给予序号作为其测试名（例如上面的#00）"}]},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"对 IDE（Goland）友好的子测试"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"有一种测试的写法是："}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"tcList := map[string][]int{\n \"t1\": {1, 2, 3},\n \"t2\": {4, 5, 9},\n}\nfor name, tc := range tcList {\n t.Run(name, func(t *testing.T) {\n require.Equal(t, tc[2], Add(tc[0], tc[1]))\n })\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"看上去没什么问题，然而有一个缺点是，这个测试对 IDE 并不友好："}]},{"type":"image","attrs":{"src":"https:\/\/static001.geekbang.org\/infoq\/79\/79c622c03b202088a1f8ebaba9a15706.png","alt":"图片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"我们无法在出错的时候对单个测试重新执行所以推荐尽可能对每个 t.Run 都要独立书写，例如："}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"f := func(a, b, exp int) func(t *testing.T) {\n return func(t *testing.T) {\n require.Equal(t, exp, Add(a, b))\n }\n}\nt.Run(\"t1\", f(1, 2, 3))\nt.Run(\"t2\", f(4, 5, 9))"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https:\/\/static001.geekbang.org\/infoq\/56\/569cad132db865846b91302381f547e5.png","alt":"图片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"测试分包"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"我们上面的 add.go 和 add_test.go 文件都处于同一个目录下，顶部的 package 名称都是 add，那么在写测试的过程中，也可以为测试启用与非测试文件不同的包名，例如我们现在将测试文件的包名改为 add_test："}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":" \/\/ add_test.go\npackage add_test\n\nimport (\n \"testing\"\n)\n\nfunc TestAdd(t *testing.T) {\n res := Add(1, 2)\n if res != 3 {\n t.Errorf(\"the result is %d instead of 3\", res)\n }\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"这个时候执行 go test 会发现"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"% go test\n# code.byted.org\/ek\/demo_test\/t03_diffpkg_test [code.byted.org\/ek\/demo_test\/t03_diffpkg.test]\n.\/add_test.go:9:9: undefined: Add\nFAIL code.byted.org\/ek\/demo_test\/t03_diffpkg [build failed]\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"由于包名变化了，我们无法再访问到 Add 函数，这个时候我们增加 import 即可："}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":" \/\/ add_test.go\npackage add_test\n\nimport (\n \"testing\"\n\n . \"code.byted.org\/ek\/demo_test\/t03_diffpkg\"\n)\n\nfunc TestAdd(t *testing.T) {\n res := Add(1, 2)\n if res != 3 {\n t.Errorf(\"the result is %d instead of 3\", res)\n }\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"我们使用上面的方式来导入包内的函数即可。但使用了这种方式后，将无法访问包内未导出的函数（以小写开头的）。"}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"测试的工具库"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"github.com\/stretchr\/testify"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"我们可以使用强大的 testify 来方便我们写测试例如上面的测试我们可以用这个库写成："}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":" \/\/ add_test.go\npackage correct\n\nimport (\n \"testing\"\n\n \"github.com\/stretchr\/testify\/require\"\n)\n\nfunc TestAdd(t *testing.T) {\n res := Add(1, 2)\n require.Equal(t, 3, res)\n\n \/*\n must := require.New(t)\n res := Add(1, 2)\n must.Equal(3, res)\n *\/\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"如果执行失败，则会在命令行看到如下输出："}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"% go test\nok code.byted.org\/ek\/demo_test\/t04_libraries\/testify\/correct 0.008s\n--- FAIL: TestAdd (0.00s)\n add_test.go:12:\n Error Trace: add_test.go:12\n Error: Not equal:\n expected: 3\n actual : -1\n Test: TestAdd\nFAIL\nFAIL code.byted.org\/ek\/demo_test\/t04_libraries\/testify\/wrong 0.009s\nFAIL\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"库提供了格式化的错误详情（堆栈、错误值、期望值等）来方便我们调试。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"github.com\/DATA-DOG\/go-sqlmock"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"对于需要测试 sql 的地方可以使用 go-sqlmock 来测试"}]},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"优点：不需要依赖数据库"}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"缺点：脱离了数据库的具体实现，所以需要写比较复杂的测试代码"}]}]}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"github.com\/golang\/mock"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"强大的对 interface 的 mock 库，例如我们要测试函数 ioutil.ReadAll"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"func ReadAll(r io.Reader) ([]byte, error)\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"我们 mock 一个 io.Reader"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"\/\/ package: 输出包名\n\/\/ destination: 输出文件\n\/\/ io: mock对象的包\n\/\/ Reader: mock对象的interface名\nmockgen -package gomock -destination mock_test.go io Reader\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"可以在目录下看到 mock_test.go 文件里，包含了一个 io.Reader 的 mock 实现我们可以使用这个实现去测试 ioutil.Reader，例如"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"ctrl := gomock.NewController(t)\ndefer ctrl.Finish()\nm := NewMockReader(ctrl)\nm.EXPECT().Read(gomock.Any()).Return(0, errors.New(\"error\"))\n_, err := ioutil.ReadAll(m)\nrequire.Error(t, err)\n"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"net\/http\/httptest"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"通常我们测试服务端代码的时候，会先启动服务，再启动测试。官方的 httptest 包给我们提供了一种方便地启动一个服务实例来测试的方法。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"其他"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"其他一些测试工具可以前往 awesome-go#testing 查找"}]},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"https:\/\/github.com\/avelino\/awesome-go#testing"}]}]}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"如何写好测试"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"上面介绍了测试的基本工具和写法，我们已经完成了“必先利其器”，下面我们将介绍如何“善其事”。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"并发测试"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在平时，大家写服务的时候，基本都必须考虑并发，我们使用 IDE 测试的时候，IDE 默认情况下并不会主动测试并发状态，那么如何保证我们写出来的代码是并发安全的？我们来举个例子，比如我们有个计数器，作用就是计数。"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"type Counter int32\n\nfunc (c *Counter) Incr() {\n *c++\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"很显然这个计数器在并发情况下是不安全的，那么我们如何写一个测试来做这个计数器的并发测试呢？"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"import (\n \"sync\"\n \"testing\"\n\n \"github.com\/stretchr\/testify\/require\"\n)\n\nfunc TestA_Incr(t *testing.T) {\n var a Counter\n eg := sync.WaitGroup{}\n count := 10\n eg.Add(count)\n for i := 0; i 0 && j > 0 {\n i += j\n }\n return i\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"下面的测试用例覆盖率达到了 100%，但是并没有测试到所有的分支"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"func TestAdd(t *testing.T) {\n res := AddIfBothPositive(1, 2)\n require.Equal(t, 3, res)\n}\n"}]},{"type":"heading","attrs":{"align":null,"level":4},"content":[{"type":"text","text":"并没有处理异常\/边界条件"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"func Divide(i, j int) int {\n return i \/ j\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Divide 函数并没有处理除数为 0 的情况，而单元测试的覆盖率是 100%"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"func TestAdd(t *testing.T) {\n res := Divide(6, 2)\n require.Equal(t, 3, res)\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"上面的例子说明 100%的测试覆盖并不是真的“100%覆盖”了所有的代码运行情况。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"覆盖率的统计方法"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"测试覆盖率的统计方法一般是：测试中执行到的代码行数 \/ 测试的代码的总行数然而代码在实际运行中，每一行运行到的概率、出错的严重程度等等也是不同的，所以我们在追求高覆盖率的同时，不能迷信覆盖率。"}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"测试是不怕重复书写的"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"这里的重复书写，可以一定程度上认为是“代码复用”的反义词。我们主要从下面的几方面来说。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"重复书写类似的测试用例"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"测试用例只要不是完全一致，那么即便是比较雷同的测试用例，我们都可以认为是有意义的，没有必要为了代码的精简特地删除，例如我们测试上面的 Add 函数"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"func TestAdd(t *testing.T) {\n t.Run(\"fixed\", func(t *testing.T) {\n res := Add(1, 2)\n require.Equal(t, 3, res)\n })\n t.Run(\"random\", func(t *testing.T) {\n a := rand.Int()\n b := rand.Int()\n res := Add(a, b)\n require.Equal(t, a+b, res)\n })\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"虽然第二个测试看起来覆盖了第一个测试，但没有必要去特地删除第一个测试，越多的测试越能增加我们代码的可靠性。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"重复书写（源）代码中的定义和逻辑"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"比如我们有一份代码"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"package add\n\nconst Value = 3\n\nfunc AddInternalValue(a int) int {\n return a + Value\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"测试为"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"func TestAdd(t *testing.T) {\n res := AddInternalValue(1)\n require.Equal(t, 1+Value, res)\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"看起来非常完美，但是如果某天内部变量 Value 的值被不小心改动了，那么这个测试无法反应出这个改动，也就无法及时发现这个错误了。如果我们写成"}]},{"type":"codeblock","attrs":{"lang":null},"content":[{"type":"text","text":"func TestAdd(t *testing.T) {\n const value = 3\n res := AddInternalValue(1)\n require.Equal(t, 1+value, res)\n}\n"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"就不用担心无法发现常量值的变化了。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"本文转载自：字节跳动技术团队（ID：toutiaotechblog）"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"原文链接："},{"type":"link","attrs":{"href":"https:\/\/mp.weixin.qq.com\/s\/9x8wdhDDMn-UBkTPG4SsCg","title":"xxx","type":null},"content":[{"type":"text","text":"一个测试工程师走进一家酒吧……"}]}]}]}