Question

考虑这段代码：

Point findIntersection(Line l1, Line l2) 
{
  int T1, T2;
  T2 = (l2.d.x*(l1.p.y-l2.p.y) + l2.d.y*(l2.p.x-l1.p.x))/(l1.d.x*l2.d.y - l1.d.y*l2.d.x);
  T1 = (l1.p.x+l1.d.x*T2-l2.p.x)/l2.d.x;
  if (T1>0 && 0<T2<1) {
    return {l2.p.x+l2.d.x*T1, l2.p.y+l2.d.y*T1};
  }
}

（完整代码http://pastebin.com/M6G40F4M）

此代码在第3行和第4行（以及较大的代码段中的13＆amp; 14）中引起浮点异常。我的问题是为什么会发生这种情况，以及找到两条线相交的正确方法是什么。我知道这些错误通常在除以零时发生，但我不确定我在做什么以及如何防止它。

Answer 1

你除以零让我们来证明一下。

我将编译警告，以及未定义的行为清理程序。

clang++-3.9 -std=c++1z -g -Weverything -fsanitize=undefined  -o main t.cpp

警告：

main.cpp:16:12: warning: generalized initializer lists are incompatible with C++98 [-Wc++98-compat]
    return {l2.p.x+l2.d.x*T1, l2.p.y+l2.d.y*T1};
           ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
main.cpp:10:7: warning: no previous prototype for function 'findIntersection' [-Wmissing-prototypes]
Point findIntersection(Line l1, Line l2)
      ^
main.cpp:18:1: warning: control may reach end of non-void function [-Wreturn-type]
}
^
main.cpp:22:20: warning: generalized initializer lists are incompatible with C++98 [-Wc++98-compat]
    case 0: return {r.p.x, r.d.y};
                   ^~~~~~~~~~~~~~
main.cpp:25:20: warning: generalized initializer lists are incompatible with C++98 [-Wc++98-compat]
    case 3: return {r.d.x, r.p.y};
                   ^~~~~~~~~~~~~~
main.cpp:19:7: warning: no previous prototype for function 'getRectPoint' [-Wmissing-prototypes]
Point getRectPoint(Rect r, int n)
      ^
main.cpp:27:1: warning: control may reach end of non-void function [-Wreturn-type]
}
^

运行时

：

main.cpp:13:57: runtime error: division by zero
[1]    26614 floating point exception (core dumped)  ./main

第13行：

T2 = (l2.d.x*(l1.p.y-l2.p.y) + l2.d.y*(l2.p.x-l1.p.x))/(l1.d.x*l2.d.y - l1.d.y*l2.d.x);

找到交集时的浮点异常

1 个答案: