關於測試，我說的其實是

關於測試我說的其實是......

Hugo @AgileCommunity.tw

先調查⼀一下

• 誰會寫 Python 程式

• 誰玩過 BDD (Behavior-driven Development)

• 誰玩過 TDD (Test-driven Development)

• 誰玩過 CI (Continuous Integration)

故事是這樣的... 有⼀一天，PM有個想法

I have a DREAM!

網站上提供計算機的服務

功能 (Feature)

網路計算機 (Web Calculator)

1+1=2雲端技術

⼯工程計算機要跟⼿手機App結合

可以產⽣生⼤大數據匯率轉換

⼀一個功能，各⾃自表述。

使⽤用者故事 (User Story)

As a student of primary school

In order to finish my homework

I want to do arithmetic operations

使⽤用者故事提供問題的脈絡1+1=2

規格書• 滿⾜足四則運算

規格書 part2• 滿⾜足四則運算

• 運算⼦子優先順序

• 交換律

• 結合律

• 分配律

規格書 part 3• 滿⾜足四則運算

• 運算⼦子優先順序: 先括號，再× ÷，後 + −

• 交換律: x∗y = y∗x ∀ x,y ∈ S

• 結合律:(x∗y)∗z = (x∗y)∗z ∀ x,y,z ∈ S

• 分配律: x∗(y+z) = (x∗y)+(x∗z) ∀ x,y,z ∈ S

圖⽚片來源 http://goo.gl/sKZQGX

http://goo.gl/sKZQGX

能不能舉例說明？

關於⾏行為驅動開發 (Behave Driven Development, BDD)

場景 (Scenario)Scenario Outline: do simple operations Given I enter <expression> When I press "=" button Then I get the answer <answer>

Examples: | expression | answer | | 3 + 2 | 5 | | 3 - 2 | 1 | | 3 * 2 | 6 | | 3 / 2 | 1.5 | | 3 +-*/ 2 | Invalid Input | | hello world | Invalid Input |

加法/乘法交換律Scenario Outline: satisfy commutative property When I enter <expression1> first And I enter <expression2> again Then I get the same answer

Examples: | expression1 | expression2 | | 3 + 4 | 4 + 3 | | 2 * 5 | 5 * 2 |

加法/乘法結合律Scenario Outline: satisfy associative property When I enter <expression1> first And I enter <expression2> again Then I get the same answer

Examples: | expression1 | expression2 | | (2 + 3) + 4 | 2 + (3 + 4) | | 2 * (3 * 4) | (2 * 3) * 4 |

乘法左/右分配律Scenario Outline: satisfy distributive property When I enter <expression1> first And I enter <expression2> again Then I get the same answer

Examples: | expression1 | expression2 | | 2 * (1 + 3) | (2*1) + (2*3) | | (1 + 3) * 2 | (1*2) + (3*2) |

RD: 為什麼不測試這個?Scenario Outline: parse an expression Given I enter <expression> When I press "=" button Then I get an <array>

Examples: | expression | array | | 1+2 | ['1','+','2'] | | 1*2 | ['1','*','2'] |


“Because designing the technical solution is not the purpose of the specification, you should focus only on writing scenarios that relate to the business rules.”

- Executable Specification with Scrum

QA: 驗收測試，讓專業的來

執⾏行測試$ python manage.py behave --dry-run…(略)

You can implement step definitions for undefined steps with these snippets:

@given(u'I enter "3+2"')def step_impl(context): raise NotImplementedError(u'STEP: Given I enter "3+2"')

@when(u'I press "=" button')def step_impl(context): raise NotImplementedError(u'STEP: When I press "=" button')

@then(u'I get the answer "5"')def step_impl(context): raise NotImplementedError(u'STEP: Then I get the answer "5"')

重構步驟@given(u'I enter {expr}')def step_impl(context, expr): raise NotImplementedError(u'STEP: Given I enter {expr}')

@when(u'I press "=" button')def step_impl(context): raise NotImplementedError(u'STEP: When I press "=" button')

@then(u'I get the answer {answer}')def step_impl(context, answer): raise NotImplementedError(u'STEP: Then I get the answer {answer}')

…(略)

複製貼上，修修改改

執⾏行測試$ python manage.py behave

Creating test database for alias 'default'...Feature: Web calculator # features/calc.feature:3 As a student In order to finish my homework I want to do arithmatical operations Scenario Outline: do simple operations -- @1.1 Given I enter 3 + 2 Traceback (most recent call last): ...(略) NotImplementedError: STEP: Given I enter {expr}

When I press "=" button Then I get the answer 5

溫馨提⽰示：還沒拿掉 NotImplementError

重構步驟from calc.calculator import Calculator

@given(u'I enter {expr}')def step_impl(context, expr): context.expr = expr

@when(u'I press "=" button')def step_impl(context): calc = Calculator() context.answer = calc.evalString(context.expr)

@then(u'I get the answer {answer}')def step_impl(context, answer): assert context.answer == answer

假裝類別存在

假裝有個⽅方法


Creating test database for alias 'default'...Exception ImportError: No module named 'calc.calculator'; 'calc' is not a packageTraceback (most recent call last): ...(略) File "/home/vagrant/myWorkspace/demo/features/steps/calc.py", line 1, in <module> from calc.calculator import CalculatorImportError: No module named 'calc.calculator'; 'calc' is not a package

溫馨提⽰示：還沒實作 calc.calculator

拉出介⾯面#file: calc/calculator.py

class Calculator:

def evalString(self, string): return 0

回傳的預設值


Creating test database for alias 'default'...Feature: Web calculator # features/calc.feature:3 As a student In order to finish my homework I want to do arithmatical operations Scenario Outline: do simple operations -- @1.1 Given I enter 3 + 2 When I press "=" button Then I get the answer 5 Traceback (most recent call last): ...(略) File "features/steps/calc.py", line 19, in step_impl assert context.answer == answer AssertionError

QA: 接下來就是 RD 的事了


通過場景轉換成驗收測試，規格變成⼀一份可執⾏行的活⽂文件。

Acceptance TestScenario

User Story

Requirement Specification

confirmsillustrates

executes

有些⼈人熱愛中華⽂文化

圖⽚片來源 http://goo.gl/BQkP82

http://goo.gl/BQkP82

說中⽂文也⾏行場景⼤大綱: 做簡單的運算假設< 我輸⼊入<expression> 當< 我按下等號按鈕那麼< 我得到的答案是<answer>

例⼦子: | expression | answer | | 3 + 2 | 5 | | 3 - 2 | 1 | | 3 * 2 | 6 | | 3 / 2 | 1.5 | | 3 +-*/ 2 | Invalid Input | | hello world | Invalid Input |

執⾏行測試$ python manage.py behave --dry-run --include zh_calc

功能: 網⾴頁計算機 # features/zh_calc.feature:2 ⾝身為⼀一個學⽣生為了完成家庭作業我想要做算術運算場景⼤大綱: 做簡單的運算 -- @1.1 假設 < 我輸⼊入3 + 2 當 < 我按下等號按鈕那麼 < 我得到的答案是5

…(略)


BDD 不只是測試框架，

更是溝通需求的哲學。

關於測試驅動開發 (Test Driven Development, TDD)

單元測試 (Unit Test)from django.test import TestCasefrom calc.calculator import Calculator

# Create your tests here.class TestCalculator(TestCase):

def setUp(self): self.calc = Calculator()

def test_evalString(self): evalString = self.calc.evalString self.assertEqual(evalString('0'), 0)

執⾏行測試$ python manage.py test -v2

test_evalString (calc.tests.TestCalculator) ... ok

--------------------------------------------------------------Ran 1 test in 0.002s

OK

該提交程式碼與測試了

記得版本控制$ git init$ git add .$ git commit -m "init project"

再多⼀一點點測試from django.test import TestCasefrom calc.calculator import Calculator



def test_evalString(self): evalString = self.calc.evalString self.assertEqual(evalString('0'), 0) self.assertEqual(evalString('1'), 1)

Baby step，每次前進⼀一⼩小步

執⾏行測試$ python manage.py test -v2...（略)

test_evalString (calc.tests.TestCalculator) ... FAIL

==============================================================FAIL: test_evalString (calc.tests.TestCalculator)--------------------------------------------------------------Traceback (most recent call last): File "/home/vagrant/myWorkspace/demo/calc/tests.py", line 13, in test_evalString self.assertEqual(evalString('1'), 1)AssertionError: 0 != 1

溫馨提⽰示：接著完善 evalString()


測試先⾏行，它會告訴你下⼀一步該怎麼⾛走。

RD: 準備開⼯工啦～

待辦清單• 數字求值 • 錯誤處理 • 處理簡單數學運算 • 處理先乘除後加減 • 處理括號運算 • 交換律、結合律、分配律 • 確認驗收測試通過

3 + 2 − 1 = (3 + 2) − 1

⇒ [[‘3’, ‘2’, ‘+’], ‘1’, ‘−’] ⇒ [‘3’, ‘2’, ‘+’, ‘1’, ‘−’]

3 + 2 × 1 = 3 + (2 × 1)

⇒ [‘3’, [‘2’, ‘1’, ‘×’], ‘+’] ⇒ [‘3’, ‘2’, ‘1’, ‘×’, ‘+’]

解析⼆二元運算

3 2 1

+

×

3 2 1

+

-

待辦清單• 數字求值：解析字串、解析 exprStack • 錯誤處理 • 處理簡單數學運算 • 處理先乘除後加減 • 處理括號運算 • 交換律、結合律、分配律 • 確認驗收測試通過

增加測試from django.test import TestCasefrom calc.calculator import Calculator



…(略)

def test_parseString(self): parseString = self.calc.parseString self.assertEqual(parseString('0'), ['0']) self.assertEqual(parseString('1'), ['1'])

新增解析字串測試


==============================================================ERROR: test_parseString (calc.tests.TestCalculator)--------------------------------------------------------------Traceback (most recent call last): File "/home/vagrant/myWorkspace/demo/calc/tests.py", line 11, in test_parseString parseString = self.calc.parseStringAttributeError: 'Calculator' object has no attribute 'parseString'

溫馨提⽰示：接著實作解析字串的⽅方法

進⾏行實作class Calculator: def __init__(self): self.exprStack = []

integer = Word(nums) self.expr = integer + StringEnd()

def parseString(self, string): self.exprStack = [] return self.expr.parseString(string).asList()

…(略)

integer :: '0'...'9'*expr :: integer

執⾏行測試$ python manage.py test -v2...(略)test_evalString (calc.tests.TestCalculator) ... FAILtest_parseString (calc.tests.TestCalculator) ... ok




…(略)

def test_evalStack(self): evalStack = self.calc.evalStack self.assertEqual(evalStack(['0']), 0) self.assertEqual(evalStack(['1']), 1)

新增從 stack 求值的測試

執⾏行測試$ python manage.py test...(略)==============================================================ERROR: test_evalStack (calc.tests.TestCalculator)--------------------------------------------------------------Traceback (most recent call last): File "/home/vagrant/myWorkspace/demo/calc/tests.py", line 16, in test_evalStack evalStack = self.calc.evalStackAttributeError: 'Calculator' object has no attribute 'evalStack'

溫馨提⽰示：接著實作從 stack 求值的⽅方法

進⾏行實作class Calculator: def __init__(self): self.exprStack = []

def pushStack(s, l, t): self.exprStack.append(t[0])

integer = Word(nums).addParseAction(pushStack) self.expr = integer + StringEnd()

def evalStack(self, stack): op = stack.pop() return float(op)

…(略)從 stack 取值後回傳

執⾏行測試$ python manage.py test -v2...(略)test_evalStack (calc.tests.TestCalculator) ... oktest_evalString (calc.tests.TestCalculator) ... FAILtest_parseString (calc.tests.TestCalculator) ... ok

沿⽤用測試from django.test import TestCasefrom calc.calculator import Calculator



…(略)

def test_evalString(self): evalString = self.calc.evalString self.assertEqual(evalString('0'), 0) self.assertEqual(evalString('1'), 1)

沿⽤用先前字串求值的測試


test_evalString (calc.tests.TestCalculator) ... FAIL

==============================================================FAIL: test_evalString (calc.tests.TestCalculator)--------------------------------------------------------------Traceback (most recent call last): File "/home/vagrant/myWorkspace/demo/calc/tests.py", line 13, in test_evalString self.assertEqual(evalString('1'), 1)AssertionError: 0 != 1

溫馨提⽰示：接著完善 evalString⽅方法

進⾏行實作class Calculator: def __init__(self): ...(略)

def parseString(self, string): ...(略)

def evalStack(self, stack): ...(略)

def evalString(self, string): self.parseString(string) return self.evalStack(self.exprStack)

介⾯面不做複雜的⼯工作

執⾏行測試$ python manage.py test -v2...(略)test_evalStack (calc.tests.TestCalculator) ... oktest_evalString (calc.tests.TestCalculator) ... oktest_parseString (calc.tests.TestCalculator) ... ok

--------------------------------------------------------------Ran 3 tests in 0.010s

OK

$ git add .$ git commit -m "test evalStack, evalString, parseString: ok"



單元測試也是程式碼的⼀一部分。




…(略)

def test_invalid_input(self): evalString = self.calc.evalString self.assertEqual(evalString('hello world'), 'Invalid Input')

你無法避免使⽤用者出怪招

執⾏行測試$ python manage.py test

==============================================================ERROR: test_invalid_input (calc.tests.TestCalculator)--------------------------------------------------------------Traceback (most recent call last): ...(略) File "/home/vagrant/myWorkspace/venv/lib/python3.5/site-packages/pyparsing.py", line 1936, in parseImpl raise ParseException(instring, loc, self.errmsg, self)pyparsing.ParseException: Expected W:(0123...) (at char 0), (line:1, col:1)

溫馨提⽰示：⺫⽬目前有看不懂的表⽰示式

進⾏行實作class Calculator:

...(略)

def evalString(self, string): try: self.parseString(string) return self.evalStack(self.exprStack) except ParseException: return 'Invalid Input'

針對無法解析的字串，預設回覆值


....--------------------------------------------------------------Ran 4 tests in 0.005s

OK

$ git add .$ git commit -m "handle ParseException"



Test

Code Refactor

Start

待辦清單• 數字求值：解析字串、解析 exprStack • 錯誤處理 • 處理簡單數學運算：解析字串、字串求值 • 處理先乘除後加減 • 處理括號運算 • 交換律、結合律、分配律 • 確認驗收測試通過




…(略)

def test_parseString(self): parseString = self.calc.parseString self.assertEqual(parseString('0'), ['0']) self.assertEqual(parseString('1'), ['1']) self.assertEqual(parseString('3+2'), ['3', '2', '+'])

Baby step，再前進⼀一⼩小步

執⾏行測試$ python manage.py test -v2

==============================================================ERROR: test_parseString (calc.tests.TestCalculator)--------------------------------------------------------------Traceback (most recent call last): File "/home/vagrant/myWorkspace/demo/calc/tests.py", line 14, in test_parseString self.assertEqual(parseString('3+2'), ['3', '2', '+']) ...(略) raise ParseException(instring, loc, self.errmsg, self)pyparsing.ParseException: Expected end of text (at char 1), (line:1, col:2)

溫馨提⽰示：⺫⽬目前有看不懂的表⽰示式


def __init__(self): self.exprStack = [] def pushStack(s, l, t): self.exprStack.append(t[0])

integer = Word(nums).addParseAction(pushStack) op = Literal('+') | Literal('-') | Literal('*') | Literal('/') expr = integer + ZeroOrMore((op + integer).addParseAction(pushStack))

self.expr = expr + StringEnd()

def parseString(self, string): self.exprStack = [] self.expr.parseString(string) return self.exprStack

integer :: '0'...'9'*op :: '+' | '-' | '*' | '/'expr :: integer [op integer]*

執⾏行測試$ python manage.py test -v2...(略)test_evalStack (calc.tests.TestCalculator) ... oktest_evalString (calc.tests.TestCalculator) ... oktest_invalid_input (calc.tests.TestCalculator) ... oktest_parseString (calc.tests.TestCalculator) ... ok

--------------------------------------------------------------Ran 4 tests in 0.007s

OK

$ git add .$ git commit -m "parseString of '3+2': ok"





…(略)

def test_num_op_num(self): evalString = self.calc.evalString self.assertEqual(evalString('3+2'), 5) self.assertEqual(evalString('3-2'), 1) self.assertEqual(evalString('3*2'), 6) self.assertEqual(evalString('3/2'), 1.5)

測試由算數表⽰示式求值


==============================================================ERROR: test_num_op_num (calc.tests.TestCalculator)--------------------------------------------------------------Traceback (most recent call last): …(略) File "/home/vagrant/myWorkspace/demo/calc/calculator.py", line 29, in evalStack return float(op)ValueError: could not convert string to float: '+'

溫馨提⽰示：無法處理加號

進⾏行實作from calc.scalc import SimpleCalculator

class Calculator:

def __init__(self): …(略)

calc = SimpleCalculator() self.opfun = { '+' : (lambda a, b: calc.add(a,b)), '-' : (lambda a, b: calc.sub(a,b)), '*' : (lambda a, b: calc.mul(a,b)), '/' : (lambda a, b: calc.div(a,b)) }

def evalStack(self, stack): op = stack.pop() if op in '+-*/': op2 = self.evalStack(stack) op1 = self.evalStack(stack) return self.opfun[op](op1, op2) else: return float(op)

使⽤用另⼀一個團隊開發的模組

呼叫處理符號的⽅方法

對應符號與處理⽅方法


==============================================================ERROR: test_num_op_num (calc.tests.TestCalculator)--------------------------------------------------------------Traceback (most recent call last): ...(略) File "/home/vagrant/myWorkspace/demo/calc/calculator.py", line 25, in <lambda> '+' : (lambda a, b: calc.add(a,b)), File "/home/vagrant/myWorkspace/demo/calc/scalc.py", line 6, in add raise NotImplementedErrorNotImplementedError

溫馨提⽰示：SimpleCalculator.add ⽅方法尚未實作

NotImplementedError?!

相依元件尚未完成#file: calc/scalc.py

class SimpleCalculator:

def add(self, a, b): raise NotImplementedError

def sub(self, a, b): raise NotImplementedError

def mul(self, a, b): raise NotImplementedError

def div(self, a, b): raise NotImplementedError

圖⽚片來源 http://goo.gl/ciEspm

http://goo.gl/ciEspm

眼前有兩條路

都是 they 的錯 It’s time to Facebook

⼩小兵 idol…

⾏行不⾏行啊

⾛走！買飲料沒事做，趕快裝忙

等

使⽤用測試替⾝身 (Test Double)

圖⽚片來源 http://goo.gl/71wmBt

http://goo.gl/71wmBt

整合測試

SUT: System Under Test DOC: Depended-on Component

Setup

Exercise

Verify

Teardown

SUT DOC

單元測試

SUT Test Double

Setup

Exercise

Verify

Teardown

測試替⾝身• 種類

• Test spy, Test stub, Mock object, Fack object, Dummy object

• ⺫⽬目的

• 不依賴其他元件，做到真正的單元測試

• 可控制的測試環境


Mock object

依賴注⼊入 (Dependency Injection, DI)

Builder Calculator

SimpleCalculator

interface add(), sub(), mul(), div()

3. use

1. create

2. inject independence


def __init__(self, calc): …(略)

calc = SimpleCalculator() self.opfun = { '+' : (lambda a, b: calc.add(a,b)), '-' : (lambda a, b: calc.sub(a,b)), '*' : (lambda a, b: calc.mul(a,b)), '/' : (lambda a, b: calc.div(a,b)) }

依賴注⼊入

修改測試class TestCalculator(TestCase):

def setUp(self): add_dict = {(3,2) : 5} sub_dict = {(3,2) : 1} mul_dict = {(3,2) : 6} div_dict = {(3,2) : 1.5}

def add(*args): return add_dict[args] def sub(*args): return sub_dict[args] def mul(*args): return mul_dict[args] def div(*args): return div_dict[args]

scalc = SimpleCalculator() scalc.add = MagicMock(side_effect = add) scalc.sub = MagicMock(side_effect = sub) scalc.mul = MagicMock(side_effect = mul) scalc.div = MagicMock(side_effect = div)

self.calc = Calculator(scalc)

...(略)

測試替⾝身

依賴注⼊入

只會回答 3+2, 3-2, 3*2, 3/2 的問題


.....--------------------------------------------------------------Ran 5 tests in 0.012s

OK

$ git add .$ git commit -m "DI and mock of SimpleCalculator"


實作細節講太多讓⼈人昏昏欲睡

接下來稍微加速⼀一下

圖⽚片來源 http://goo.gl/IpX6lw

http://goo.gl/IpX6lw


Mock object

增加測試class TestCalculator(TestCase):

def setUp(self): add_dict = {…} sub_dict = {…} mul_dict = {…} div_dict = {…} …(略)

def test_order_of_operations(self): evalString = self.calc.evalString self.assertEqual(evalString('4+3*2'), 10) self.assertEqual(evalString('9-3*2+2/1'), 5)

增加偽造的範圍

測試先乘除、後加減

執⾏行測試$ python manage.py test...(略)==============================================================FAIL: test_order_of_operations (calc.tests.TestCalculator)--------------------------------------------------------------Traceback (most recent call last): File "/home/vagrant/myWorkspace/demo/calc/tests.py", line 62, in test_order_of_operations self.assertEqual(evalString('4+3*2'), 10)AssertionError: 14.0 != 10

溫馨提⽰示: 還沒處理“先乘除、後加減”


def __init__(self, calc): self.exprStack = [] def pushStack(s, l, t): self.exprStack.append(t[0])

integer = Word(nums).addParseAction(pushStack) addop = Literal('+') | Literal('-') mulop = Literal('*') | Literal('/')

atom = integer term = atom + ZeroOrMore((mulop + atom).addParseAction(pushStack)) expr = term + ZeroOrMore((addop + term).addParseAction(pushStack)) self.expr = expr + StringEnd()

...(略) integer :: '0'...'9'*addop :: '+' | '-'mulop :: '*' | '/'atom :: integerterm :: atom [mulop atom]*expr :: term [addop term]*


......--------------------------------------------------------------Ran 6 tests in 0.018s

OK

$ git add .$ git commit -m "evalString can handle the order of operations"



Mock object


def setUp(self): add_dict = {…} sub_dict = {…} mul_dict = {…} div_dict = {…} …(略)

def test_parentheses(self): evalString = self.calc.evalString self.assertEqual(evalString('(4+3)*2'), 14) self.assertEqual(evalString('(9-3)*(2+2)/1'), 24)

增加偽造的範圍

測試括號運算

執⾏行測試$ python manage.py test...(略)==============================================================FAIL: test_parentheses (calc.tests.TestCalculator)--------------------------------------------------------------Traceback (most recent call last): File "/home/vagrant/myWorkspace/demo/calc/tests.py", line 67, in test_parentheses self.assertEqual(evalString('(4+3)*2'), 14)AssertionError: 'Invalid Input' != 14

溫馨提⽰示: 還沒處理括號運算


def __init__(self, calc): ...(略)

integer = Word(nums).addParseAction(pushStack) addop = Literal('+') | Literal('-') mulop = Literal('*') | Literal('/') lpar = Literal('(') rpar = Literal(')')

expr = Forward() atom = integer | lpar + expr + rpar term = atom + ZeroOrMore((mulop + atom).addParseAction(pushStack)) expr << term + ZeroOrMore((addop + term).addParseAction(pushStack)) self.expr = expr + StringEnd()

...(略)integer :: '0'...'9'*addop :: '+' | '-'mulop :: '*' | '/'atom :: integer | '(' + expr + ')'term :: atom [mulop atom]*expr :: term [addop term]*


.......--------------------------------------------------------------Ran 7 tests in 0.015s

OK

$ git add .$ git commit -m "evalString can handle parentheses"



Mock object

好消息! SimpleCalculator 完成

相依元件已經完成#file: calc/scalc.py

class SimpleCalculator:

def add(self, a, b): return a+b

def sub(self, a, b): return a-b

def mul(self, a, b): return a*b

def div(self, a, b): return a/b


Mock object

換上 SimpleCalculator

修改測試class TestCalculator(TestCase):

def setUp(self): """ add_dict = {…} sub_dict = {…} mul_dict = {…} div_dict = {…}

def add(*args): return add_dict[args] def sub(*args): return sub_dict[args] def mul(*args): return mul_dict[args] def div(*args): return div_dict[args]

scalc = SimpleCalculator() scalc.add = MagicMock(side_effect = add) scalc.sub = MagicMock(side_effect = sub) scalc.mul = MagicMock(side_effect = mul) scalc.div = MagicMock(side_effect = div) """ self.calc = Calculator()

...(略)

註解程式碼

不傳 mock object 進去

修改實作from calc.scalc import SimpleCalculator...(略)

class Calculator:

def __init__(self, calc = SimpleCalculator()): self.exprStack = [] def pushStack(s, l, t): self.exprStack.append(t[0])

…(略)

預設使⽤用 SimpleCalculator


.......--------------------------------------------------------------Ran 7 tests in 0.015s

OK

$ git add .$ git commit -m "use SimpleCalculator instead of mock object"



Mock object



…(略)

def test_commutative_property(self): evalString = self.calc.evalString self.assertEqual(evalString('3+4'), evalString('4+3')) self.assertEqual(evalString('2*5'), evalString('5*2'))

def test_associative_property(self): evalString = self.calc.evalString self.assertEqual(evalString('(5+2) + 1'), evalString('5 + (2+1)')) self.assertEqual(evalString('(5*2) * 3'), evalString('5 * (2*3)'))

def test_distributive_property(self): evalString = self.calc.evalString self.assertEqual(evalString('2 * (1+3)'), evalString('(2*1) + (2*3)')) self.assertEqual(evalString('(1+3) * 2'), evalString('(1*2) + (3*2)'))

分配律測試

結合律測試

交換律測試


..........--------------------------------------------------------------Ran 10 tests in 0.021s

OK

$ git add .$ git commit -m "satisfy commutative, associative, distributive properties"



“I get paid for code that works, not for tests, so my philosophy is to

test as little as possible to reach a given level of confidence.”

- Kent Beck’s answer to How deep are your unit tests?

http://stackoverflow.com/questions/153234/how-deep-are-your-unit-tests


Mock object



Creating test database for alias 'default'...Feature: Web calculator # features/calc.feature:3 As a student In order to finish my homework I want to do arithmatical operations Scenario Outline: do simple operations -- @1.1 Given I enter 3 + 2 When I press "=" button Then I get the answer 5

…(略)

1 feature passed, 0 failed, 0 skipped12 scenarios passed, 0 failed, 0 skipped36 steps passed, 0 failed, 0 skipped, 0 undefined


各層測試的⺫⽬目不⼀一樣

User Story

Design Unit Tests

Acceptance Tests

make sure you do things right

make sure you do right things


Mock object


測試覆蓋率

測試成功趨勢

靜態檢查結果

專案健康狀態


讓⾃自動化測試成為⼀一張開發的安全網

醜媳婦⾒見公婆

圖⽚片來源 http://goo.gl/yMDEQd

http://goo.gl/yMDEQd

Templates

Views

Browser

Models

Database

URLs

Django MTV

Template<form ation="." method="post"> {% csrf_token %} <input id="expr" type="text" name="expr" value="{{ value }}"> <input type="submit" value="="></form>

Viewfrom django.shortcuts import renderfrom django.http import HttpResponsefrom .calculator import Calculator

# Create your views here.def calc(request): value = ''

if request.method == 'POST': calc = Calculator() expr = request.POST['expr'] value = calc.evalString(expr)

return render(request, 'calculator.html', {'value': value})

這層越薄越好

URLurlpatterns = [ ...(略) url(r'^$', calc_views.calc),]


圖⽚片來源 http://goo.gl/Jgmgjy

透過 Demo 測試、回饋、學習

http://goo.gl/Jgmgjy

提交版本$ git add .$ git commit -m "add template/view/URL of web calculator"

還有哪些測試？

Automated

Functional Acceptance Tests

Manual

Showcases Usability Testing

Exploratory Testing

Unit Tests Component Tests

System Tests

Automated

Nonfunctional Acceptance Tests

(Capacity, Security, Availability…)

Manual/Automated

Testing quadrant diagram

Supp

ortin

g pr

ogr

amm

ing

Critique project

Business facing

Technology facing

https://continuousdelivery.com/foundations/test-automation/

探索式測試 (Exploratory Testing)

• “所謂探索式測試, 就是同時進⾏行分析系統, 學習系統, 設計測試, 執⾏行測試等動作. 因為⼀一開始對受測系統不太懂, 無法開始就設計到位, 需要先執⾏行⼀一下系統, 了解他是什麼, 同時也思考要如何規劃設計. 因此, 這幾個動作是交錯在⼀一起進⾏行的. 這就像敏捷開發⼀一樣, 設計, 開發和測試會同時發⽣生, 不應該分成不同階段.” - David Ko

• 除以零會爆炸、不⽀支援⼩小數點、不⽀支援負數…


唯有⾃自動化才能把⼈人⼒力從瑣碎的⼿手動測試解放出來，

去做有價值的⼯工作。

Live Demo

DOC

RecapUser Stories

As a …, I want …, So that …

Scenarios

Given … When … Then …

Steps

@given(…) def step_impl(context, …): … @when(…) def step_impl(context, …): … @then(…) def step_impl(context, …): …

Interfacebetween the delegated tasks and the domain model

Domain Models

Production code here…

Database Filesystem

Network 3rd party library

Unit Tests

Test code here…

Test Double

Spy, Stub, Mock…

PM

PM QA RD

QA RD

Hardware Remote service


只要有⼼心，⼈人⼈人都可以玩測試。

學習筆記放在 https://github.com/hugolu/learn-test

Any question, please send me pull requests :)

https://github.com/hugolu/learn-test

關於測試，我說的其實是

Education