我想有人可能会喜欢用(注释的)例子来解释:

下面，如果我的deepClone()函数正确地完成了它的工作，测试(如'it()'调用中所述)将成功:

describe('deepClone() array copy', ()=>{
    let source:any = {}
    let clone:any = source
    beforeAll(()=>{
        source.a = [1,'string literal',{x:10, obj:{y:4}}]
        clone = Utils.deepClone(source) // THE CLONING ACT TO BE TESTED - lets see it it does it right.
    })
    it('should create a clone which has unique identity, but equal values as the source object',()=>{
        expect(source !== clone).toBe(true) // If we have different object instances...
        expect(source).not.toBe(clone) // <= synonymous to the above. Will fail if: you remove the '.not', and if: the two being compared are indeed different objects.
        expect(source).toEqual(clone) // ...that hold same values, all tests will succeed.
    })
})

当然，这不是我的deepClone()的完整测试套件，因为我没有在这里测试数组中的对象文字(以及其中嵌套的对象)是否也具有不同的标识但相同的值。

我认为toEqual是检查deep equal, toBe是2个变量的相同引用

  it('test me', () => {
    expect([] === []).toEqual(false) // true
    expect([] == []).toEqual(false) // true

    expect([]).toEqual([]); // true // deep check
    expect([]).toBe([]); // false
  })

查看Jasmine源代码可以更清楚地了解这个问题。

toBe非常简单，只使用恒等/严格等式运算符===:

  function(actual, expected) {
    return {
      pass: actual === expected
    };
  }

另一方面，toEqual有近150行长，并且对内置对象有特殊处理，比如String、Number、Boolean、Date、Error、Element和RegExp。对于其他对象，它递归地比较属性。

这与相等运算符==的行为非常不同。例如:

var simpleObject = {foo: 'bar'};
expect(simpleObject).toEqual({foo: 'bar'}); //true
simpleObject == {foo: 'bar'}; //false

var castableObject = {toString: function(){return 'bar'}};
expect(castableObject).toEqual('bar'); //false
castableObject == 'bar'; //true

引用jasmine github项目，

期望(x) .toEqual (y);比较对象或原语x和y和 如果它们相等，则通过 期望(x) .toBe (y);比较对象或原语x和y并传递 如果它们是同一个物体

toEqual()比较的是原语的值，还是对象的内容。 toBe()比较引用。

以下代码/套件应该是自解释的:

describe('Understanding toBe vs toEqual', () => {
  let obj1, obj2, obj3;

  beforeEach(() => {
    obj1 = {
      a: 1,
      b: 'some string',
      c: true
    };

    obj2 = {
      a: 1,
      b: 'some string',
      c: true
    };

    obj3 = obj1;
  });

  afterEach(() => {
    obj1 = null;
    obj2 = null;
    obj3 = null;
  });

  it('Obj1 === Obj2', () => {
    expect(obj1).toEqual(obj2);
  });

  it('Obj1 === Obj3', () => {
    expect(obj1).toEqual(obj3);
  });

  it('Obj1 !=> Obj2', () => {
    expect(obj1).not.toBe(obj2);
  });

  it('Obj1 ==> Obj3', () => {
    expect(obj1).toBe(obj3);
  });
});

我想有人可能会喜欢用(注释的)例子来解释:

下面，如果我的deepClone()函数正确地完成了它的工作，测试(如'it()'调用中所述)将成功:

describe('deepClone() array copy', ()=>{
    let source:any = {}
    let clone:any = source
    beforeAll(()=>{
        source.a = [1,'string literal',{x:10, obj:{y:4}}]
        clone = Utils.deepClone(source) // THE CLONING ACT TO BE TESTED - lets see it it does it right.
    })
    it('should create a clone which has unique identity, but equal values as the source object',()=>{
        expect(source !== clone).toBe(true) // If we have different object instances...
        expect(source).not.toBe(clone) // <= synonymous to the above. Will fail if: you remove the '.not', and if: the two being compared are indeed different objects.
        expect(source).toEqual(clone) // ...that hold same values, all tests will succeed.
    })
})

当然，这不是我的deepClone()的完整测试套件，因为我没有在这里测试数组中的对象文字(以及其中嵌套的对象)是否也具有不同的标识但相同的值。