This is an interactive problem. The jury has a permutation (p) of length (n) and wants you to guess it. For this, the jury created another permutation (q) of length (n). Initially, (q) is an identity permutation ((q_i = i) for all (i)). You can ask queries to get (q_i) for any (i) you want. After each query, the jury will change (q) in the following way: At first, the jury will create a new permutation (q') of length (n) such that (q'_i = q_{p_i}) for all (i). Then the jury will replace permutation (q) with pemutation (q'). You can make no more than (2n) queries in order to quess (p). The first line of input contains a single integer (t) ((1 \leq t \leq 1000)) — the number of test cases. Interaction in each test case starts after reading the single integer (n) ((1 \leq n \leq 10^4)) — the length of permutations (p) and (q). To get the value of (q_i), output the query in the format (?) (i) ((1 \leq i \leq n)). After that you will receive the value of (q_i). You can make at most (2n) queries. After the incorrect query you will receive (0) and you should exit immediately to get Wrong answer verdict. When you will be ready to determine (p), output (p) in format (!) (p_1) (p_2) (\ldots) (p_n). After this you should go to the next test case or exit if it was the last test case. Printing the permutation is not counted as one of (2n) queries. After printing a query do not forget to output end of line and flush the output. Otherwise, you will get Idleness limit exceeded . To do this, use: fflush(stdout) or cout.flush() in C++; System.out.flush() in Java; flush(output) in Pascal; stdout.flush() in Python; see documentation for other languages. It is guaranteed that the sum of (n) over all test cases doesn't exceed (10^4). The interactor is not adaptive in this problem. Hacks: To hack, use the following fo