Question

我喜欢为包含一些浮点值的结构编写匹配器：

struct Point3D
{
float x;
float y;
float z;
}

class Interface
{
virtual void SetPoint(Point3D point) = 0;
}

class MockInterface:
    public Interface
{
   MOCK_METHOD1(SetPoint, void(Point3D point));
}

MATCHER_P(Point3DEq, point, "Comparison of a Point3D")
{
   return point.x == arg.x && point.y == arg.y && point.z == arg.z;
}

TEST(Point3DComparison, TestIfPoint3DAreEqual)
{
   MockInterface interfaceFake;
   Point setPoint = { 1.0, 1.0, 1.0 }
   EXPECT_CALL(interfaceFake, SetPoint(Point3DEq(setPoint)).WillOnce(Return());
}

我不喜欢这个Matcher，因为他并不尊重良好的浮动比较。我想用gmock或gtest的浮动比较来编写一个Matcher。看起来应该类似。

MATCHER_P(Point3DEq, point, "Comparison of a Point3D")
{
   return EXPECT_THAT(point.x, FloatEq(arg.x) && EXPECT_THAT(point.y, FloatEq(arg.y) && EXPECT_THAT(point.z, FloatEq(arg.z);
}

问题是EXPECT_THAT没有返回任何布尔值。使用gmock和gtest的功能有没有一个很好的干净方法？

Answer 1

好的方法是忘记MATCHER_P ...

使用::testing::AllOf - 并使用gmock中的正确匹配器 - 如提到的FloatEq/DoubleEq - 请参阅：

auto Point3DEq(Point3D const& p_expected) { using namespace testing; return AllOf(Field(&Point3D::x, DoubleEq(p_expected.x)), Field(&Point3D::y, DoubleEq(p_expected.y)), Field(&Point3D::z, DoubleEq(p_expected.z))); }

在Matcher中使用gtest float比较

1 个答案: