to clarify: there is no problem to use doubles for computations, but you must input data as ints and convert it to doubles (10x faster), e.g.,

using D = double;
using V = array<D,2>;
V rv(){ int x,y; cin >> x >> y; return V{D(x),D(y)}; }

Now >= 4 digits is OK.

First brushstroke has a point of turn in (0, 1), second - in (2, 1), thrid - in (1, -1), fourth - in (2.5, 0.5).
Hope, now you will understand.

There is my code, I still have WA #8:

#include <iostream>
#include <cstdio>
#include <cmath>
using namespace std;
int main()
{
    double x1, x2, y1, y2, Right = -1e9, Left = 1e9, Up = -1e9, Down = 1e9;

    double x3, y3, x0, y0, n, m;

    int N = 0;

    while (cin >> x1 >> y1 >> x2 >> y2)
    {
        N ++;
        x0 = (x1 + x2) / 2.0;
        y0 = (y1 + y2) / 2.0;

        n = x1 - x0;
        m = y1 - y0;
        x3 = m + x0;
        y3 = -n + y0;

        Left = min(Left, min(x1, min(x2, x3)));
        Right = max(Right, max(x1, max(x2, x3)));
        Up = max(Up, max(y1, max(y2, y3)));
        Down = min(Down, min(y1, min(y2, y3)));

    }

    if (N == 0) printf("%0.4f %0.4f", 0, 0);
    else printf("%0.4f %0.4f", fabs(Right - Left), fabs(Up - Down));

    return 0;
}

I don't know
But don't say taboo words!!!!!!!!!!!!!!!!!!!!!!

Please, give me answer, don't ignore my request...((

while (cin >> x1 >> ... >> ...)
{

}

I have problems with test 17. When I use standart min() and max() functions i get TLE, and when I use

if (x<xmin)
    xmin = x;

instead I get WA 17 !!! It is really mysterious.

Check empty input. (answer should be 0.0000 0.0000)