You are given an array of (n) non-negative integers (a_1, a_2, \dots, a_n). Your task is to cut it into three non-empty parts: a prefix, a middle part, and a suffix. Formally, you need to choose two integers (l) and (r) such that (1 \le l < r < n), and obtain three parts: the prefix up the element at index (l) inclusive (i.e., (a_1, a_2, \dots, a_l)); the central part from the element at index (l+1) to the element at index (r) inclusive (i.e., (a_{l+1}, a_{l+2}, \dots, a_r)); the suffix from the element at index (r+1) to (n) inclusive (i.e., (a_{r+1}, a_{r+2}, \dots, a_n)). Let (s_1, s_2, s_3) be the remainders of the sums of these parts modulo (3). In other words: (s_1 = (\sum\limits_{i=1}^{l} a_i) \bmod 3); (s_2 = (\sum\limits_{i=l+1}^{r} a_i) \bmod 3); (s_3 = (\sum\limits_{i=r+1}^{n} a_i) \bmod 3). Your task is to find such boundaries (l) and (r) that either all numbers (s_1, s_2, s_3) are different, or all numbers (s_1, s_2, s_3) are the same. The first line contains a single integer (t) ((1 \le t \le 1000)) — the number of test cases. Each test case consists of two lines: the first line contains a single integer (n) ((3 \le n \le 40)); the second line contains (n) integers (a_1, a_2, \dots, a_n) ((0 \le a_i \le 40)). For each test case, if a suitable pair of integers (l) and (r) ((1 \le l < r < n)) exists, output these two integers (if there are multiple suitable pairs, you can output any of them). Otherwise, output two integers equal to (0). Consider the examples from the statement: in the first example, the array is cut into parts (1, 2, 3), (4, 5), (6); (s_1 = s_2 = s_3 = 0); in the second example, there is no suitable cut; in the third example, the array is cut into parts (2), (1), (0); (s_1 = 2), (s_2 = 1), (s_3 = 0); in the fourth example, the array is cut