std ::在复合类错误上累积

Question

因此，我试图在std :: vector上实现一个仿函数，以对两个Point组件之一/两者执行std :: accumulate。

uint32_t EllipseFit::getSum(SampleSumTypes method)
    {
        uint32_t sum=0;

        /* determine which Sum function to use */
        switch(method)
        {
        case SUM_X:
            sum = accumulate(m_dataPoints.begin(), m_dataPoints.end(), 0, SumX);
        break;

        case SUM_Y:
            sum = accumulate(m_dataPoints.begin(), m_dataPoints.end(), 0, 
                                 [&](const Point &p) {sum += p.Y();} );
            break;

          /* similar case statements... */
        }

        return sum;
    }

因此，我正在尝试两种方法来执行“混合”累加器功能

1. 使用函子结构

/* defined a little earlier in the *.cpp file, with global storage */
    struct F1
    {
        uint32_t operator() (uint32_t sum, const Point &p)
        {
            return sum + p.X();
        }
    } SumX;

2. 具有lambda函数（请参阅SUM_Y情况）；

在两种情况下，我都得到此错误输出：

In file included from /usr/include/c++/5/algorithm:62:0,
                 from ellipse-fitting.cpp:17:
/usr/include/c++/5/bits/stl_algo.h: In instantiation of ‘_OIter std::transform(_IIter, _IIter, _OIter, _UnaryOperation) [with _IIter = __gnu_cxx::__normal_iterator<Recognition::Point*, std::vector<Recognition::Point> >; _OIter = __gnu_cxx::__normal_iterator<Recognition::Point*, std::vector<Recognition::Point> >; _UnaryOperation = Recognition::EllipseFit::calcEllipse()::<lambda(Recognition::Point&)>]’:
ellipse-fitting.cpp:168:60:   required from here
/usr/include/c++/5/bits/stl_algo.h:4175:12: error: no match for ‘operator=’ (operand types are ‘Recognition::Point’ and ‘void’)
  *__result = __unary_op(*__first);
            ^
In file included from /media/sf_shared-data/src/ml-algorithm/cpp/libimgproc/include/libimgproc.hpp:22:0,
                 from ellipse-fitting.hpp:17,
                 from ellipse-fitting.cpp:19:
/media/sf_shared-data/src/ml-algorithm/cpp/libimgproc/include/rec_types.hpp:39:11: note: candidate: Recognition::Point& Recognition::Point::operator=(const Recognition::Point&)
     class Point
           ^
/media/sf_shared-data/src/ml-algorithm/cpp/libimgproc/include/rec_types.hpp:39:11: note:   no known conversion for argument 1 from ‘void’ to ‘const Recognition::Point&’
/media/sf_shared-data/src/ml-algorithm/cpp/libimgproc/include/rec_types.hpp:39:11: note: candidate: Recognition::Point& Recognition::Point::operator=(Recognition::Point&&)
/media/sf_shared-data/src/ml-algorithm/cpp/libimgproc/include/rec_types.hpp:39:11: note:   no known conversion for argument 1 from ‘void’ to ‘Recognition::Point&&’
Makefile:56: recipe for target 'ellipse-fitting.o' failed
make[1]: *** [ellipse-fitting.o] Error 1
make[1]: Leaving directory '/media/sf_shared-data/src/ml-algorithm/cpp/libimgproc/algorithms/ellipse-fit'
Makefile:39: recipe for target 'algorithms/ellipse-fit' failed

我的猜测是我的'Point'类缺少一些东西来放置编译器，但我很困惑。

这是我的积分班：

    class Point
    {
    protected:
        int16_t px;
        int16_t py;

    public:

        Point(int16_t x=SHRT_MIN, int16_t y=SHRT_MIN);
        int16_t &X();
        int16_t X() const;

        int16_t &Y();
        int16_t Y() const;

        operator bool() const;
        bool operator !() const;
        bool operator != (const Point &pt2) const;
        bool operator <  (const Point &pt2) const;
        Point &operator -= (const Point &p2);
        double operator *(const Point &p2) const;

        bool horizontalTo(Point &pt2);
        bool verticalTo(Point &pt2);
        double dot(Point &pt2);

        friend std::ostream &operator << (std::ostream &out, Point &p);
    };

    double EllipseFit::getMeanX()
    {
        double mean = (double)getSum(SUM_X);
        return mean / m_dataPoints.size();
    }

    double EllipseFit::getMeanY()
    {
        double mean = (double)getSum(SUM_Y);
        return mean / m_dataPoints.size();
    }

    Ellipse EllipseFit::calcEllipse()
    {
        uint16_t meanX=(uint16_t)getMeanX(), meanY=(uint16_t)getMeanY();
        uint32_t A, B, C, D, E, F, G;

        /* shift the points into (U, V) points := {X - mean(X), Y - mean(Y)} */
        transform(m_dataPoints.begin(), m_dataPoints.end(), m_dataPoints.begin(), \
                [&](Point &p) {p -= Point(meanX, meanY); } );

        A=getSum(this->SUM_XX);
        B=getSum(this->SUM_XY);
        C=getSum(this->SUM_YY);
        D=getSum(this->SUM_XXX);
        E=getSum(this->SUM_XYY);
        F=getSum(this->SUM_YYY);
        G=getSum(this->SUM_XXY);

        return Ellipse(A, B, C, D, E, F, G, m_dataPoints.size(), meanX, meanY);
    }

谢谢您的建议！查尔斯

Answer 1

感谢Igor向我指出错误！转换点时，我忘记添加return语句。