B'You are given an integer array a_1, ldots, a_n , where 1 <= a_i <= n for all i . There 's a "replace" function f which takes a pair of integers (l, r) , where l <= r , as input and outputs the pair f big( (l, r) big)= <= ft( min {a_l,a_{l+1}, ldots,a_r }, , max {a_l,a_{l+1}, ldots,a_r } right). Consider repeated calls of this function. That is, from a starting pair (l, r) we get f big((l, r) big) , then f big(f big((l, r) big) big) , then f big(f big(f big((l, r) big) big) big) , and so on. Now you need to answer q queries. For the i -th query you have two integers l_i and r_i ( 1 <= l_i <= r_i <= n ). You must answer the minimum number of times you must apply the "replace" function to the pair (l_i,r_i) to get (1, n) , or report that it is impossible. The first line contains two positive integers n , q ( 1 <= n,q <= 10^5 ) -- the length of the sequence a and the number of the queries. The second line contains n positive integers a_1,a_2, ldots,a_n ( 1 <= a_i <= n ) -- the sequence a . Each line of the following q lines contains two integers l_i , r_i ( 1 <= l_i <= r_i <= n ) -- the queries. For each query, output the required number of times, or -1 if it is impossible. In the first example, n=5 and a=[2,5,4,1,3] . For the first query: (4,4) to(1,1) to(2,2) to(5,5) to(3,3) to(4,4) to ldots , so it 's impossible to get (1,5) . For the second query, you already have (1,5) . For the third query: (1,4) to(1,5) . For the fourth query: (3,5) to(1,4) to(1,5) . For the fifth query: (4,5) to(1,3) to(2,5) to(1,5) . For the sixth query: (2,3) to(4,5) to(1,3) to(2,5) to(1,5) . '...