用了ac的代码试了下，也不行，蛋都碎了！！！

我已经神经质地把所有int8换成int64了，下面代码如有bug，求狠拍！
---------------------------------------


#include <stdio.h>
#include <algorithm>

using namespace std;

void sum_up(long long *ary, long long len, long long *sum)
{
    sum[0] = ary[0];
    for(long long i = 0; i < len; i ++)
        sum[i + 1] = sum[i] + ary[i + 1];
}

bool validate(long long *ary, long long *sum, long long ary_len, long long frag_cnt, long long len)
{
    if(frag_cnt <= 0)
        return true;

    if(ary_len == 0 && frag_cnt != 0)
        return false; 

    if(sum[ary_len - 1] < frag_cnt * len)
        return false;

    long long consume = ary[ary_len - 1] / len;

    if(consume == 0)
        return false;

    return validate(ary, sum, ary_len - 1, frag_cnt - consume, len);
}


long long bin(long long *ary, long long *sum, long long ary_len, long long frag_cnt, long long begin, long long end)
{
    if(end == begin)
        return begin;

	if(end - begin == 1)
	{
		if(validate(ary, sum, ary_len, frag_cnt, end))
			return end;
		return begin;
	}

	long long middle = (begin + end) / 2;

	if(validate(ary, sum, ary_len, frag_cnt, middle))
		return bin(ary, sum, ary_len, frag_cnt, middle, end);
	else
		return bin(ary, sum, ary_len, frag_cnt, begin, middle);
}

void calc(long long *ary, long long *sum, long long ary_len, long long frag_cnt)
{

	if(!validate(ary, sum, ary_len, frag_cnt, 1))
	{
		printf("0.00\n");
		return;
	}

	long long max = sum[ary_len - 1] / frag_cnt;
	max = max > ary[ary_len - 1] ? ary[ary_len - 1] : max;

	long long result = bin(ary, sum, ary_len, frag_cnt, 1, max);

	double f = result / 100.0f;
	printf("%.2lf\n", f);
}

int main(void)
{
	long long n, k;
	long long cables[10002];
	long long sum[10002];

    while(scanf("%lld %lld", &n, &k) != EOF)
    {
        double temp;
        for(long long i = 0; i < n; i ++)
        {
            scanf("%lf", &temp);
            cables[i] = (long long)((temp + 0.005) * 100);
        }
        sort(cables, cables + n);

        sum_up(cables, n, sum);

        calc(cables, sum, n, k);
    }
    return 0;
}

• Re:回去查了下不同版本的g++的浮点处理不一样，不知道是不是这个原因 (2375) ixor 2013-04-03 19:49:04