CodeForces

B. Problems for Round

time limit per test

2 seconds

memory limit per test

256 megabytes

input

standard input

output

standard output

There are n problems prepared for the next Codeforces round. They are arranged in ascending order by their difficulty, and no two problems have the same difficulty. Moreover, there are m pairs of similar problems. Authors want to split problems between two division according to the following rules:

• Problemset of each division should be non-empty.
• Each problem should be used in exactly one division (yes, it is unusual requirement).
• Each problem used in division 1 should be harder than any problem used in division 2.
• If two problems are similar, they should be used in different divisions.

Your goal is count the number of ways to split problem between two divisions and satisfy all the rules. Two ways to split problems are considered to be different if there is at least one problem that belongs to division 1 in one of them and to division 2 in the other.

Note, that the relation of similarity is not transitive. That is, if problem i is similar to problem j and problem j is similar to problem k, it doesn't follow that i is similar to k.

Input

The first line of the input contains two integers n and m (2 ≤ n ≤ 100 000, 0 ≤ m ≤ 100 000) — the number of problems prepared for the round and the number of pairs of similar problems, respectively.

Each of the following m lines contains a pair of similar problems ui and vi (1 ≤ ui, vi ≤ n, ui ≠ vi). It's guaranteed, that no pair of problems meets twice in the input.

Output

Print one integer — the number of ways to split problems in two divisions.

Examples

input

5 21 45 2

output

2

input

3 31 22 31 3

output

0

input

3 23 13 2

output

1

Note

In the first sample, problems 1 and 2 should be used in division 2, while problems 4 and 5 in division 1. Problem 3 may be used either in division 1 or in division 2.

In the second sample, all pairs of problems are similar and there is no way to split problem between two divisions without breaking any rules.

Third sample reminds you that the similarity relation is not transitive. Problem 3 is similar to both 1 and 2, but 1 is not similar to 2, so they may be used together.

题意：cf出题，要求Div1的题要比Div2的题难，且相似的题目不能出在同以套题里。给出n道题，数字越大代表题目难度越大，给出m个相似关系。问能出多少种题。

思路：根据给出的关系找出Div1的最小数和Div2的最大数。

#include<stdio.h>#include<algorithm>using namespace std;int main(){    int n,m,MAX,MIN;    while(scanf("%d%d",&n,&m)!=EOF)    {        int MAX=1;        int MIN=n;        while(m--)        {            int x,y;            scanf("%d%d",&x,&y);            MAX=max(MAX,min(x,y));            MIN=min(MIN,max(x,y));        }        if(MIN-MAX>0)            printf("%d\n",MIN-MAX);//先把未分配的数都分到Div1 然后从Div1从小到大往Div2拿。        else            printf("0\n");    }}