I've already solved this problem in Pascal, but now I'm trying to get through with my C++ code. I see no difference to my accepted solution and interested, what the point.

Can someone help me to understand?

Compiled with Gnu C++ compiler.
#include <cstdio>
#include <vector>

#define ___MAXN 1000

typedef unsigned int uint;
std::vector< std::vector<int> > G(0);
int col[___MAXN] = {0};

void dfs(uint u, uint c)  {
    col[u] = c;
    for (uint v=0; v<G[u].size(); v++)
        if (!col[G[u][v]])
            dfs(G[u][v], 3-col[u]);
}

void colorize() {
    for (uint u=0; u<G.size(); u++)
        if (!col[u])
            dfs(u, 1);
}

int main()  {
    uint n;
    scanf("%u", &n);
    G.resize(n);

    for (uint i=0; i<n; i++) {
        static uint x;
        scanf("%u", &x);
        do  {
            G[i].push_back(x-1);
            scanf("%u", &x);
        }   while (x);
    }
    colorize();
    n = 0;
    for (uint i=0; i<G.size(); i++)
        switch (col[i]) {
        case 0:
            printf("0\n");
            return 0;
        case 1:
            n++;
            break;
        }
    printf("%u\n", n);
    char c = 0;
    for (uint i=0; i<G.size(); i++)
        if (col[i]==1)  {
            printf("%c%u", c, i+1);
            c = ' ';
        }
}