Alien Gene Editing Minimization

An alien civilization, HD1048576d, uses a gene editing technique on their life forms. Their genetic material is represented as a sequence of positive integers (each integer represents a noicleobase). Their editing process is as follows:

1. Select an editing region [l, r] in the sequence, i.e. the segment s_l, s_l+1, ... , s_r will be replaced.
2. Choose a pair of indices (i, j) with 1 ≤ i < l and r < j ≤ n so that the subsequence s_i, s_i+1, ... , s_j is processed. In the new sequence this segment becomes s_i, s_i+1, ... , s_l-1, t₁, t₂, ... , t_k, s_r+1, ... , s_j, where t₁, ... , t_k is newly produced.
3. After cutting out s_i ... s_j, the processed segment is reattached to form the final target gene sequence.

For the operation to be specific, the chosen pair of indices (i, j) must satisfy two key conditions:

• Editing gap: There must exist positions l and r with i < l and r < j. This implies that j - i ≥ 2.
• Uniqueness of recognition: The pair of noicleobases at the ends, i.e. (s_i, s_j), must be unique in the whole sequence. More formally, if there exists any other ordered pair of indices (i', j') (with i' < j') such that s_i' = s_i and s_j' = s_j, then the pair (s_i, s_j) is not suitable. Also, note that s_i ≠ s_j must hold.

The cost of editing is proportional to the length of the processed segment s_i ... s_j, that is, its length j - i + 1. The goal is to minimize this length among all valid pairs (i, j) satisfying the conditions above. If no such pair exists, output -1.

Input and Output in summary:

Given a gene sequence of length n represented by s₁, s₂, ... , s_n, find the minimal length j - i + 1 (with j - i ≥ 2) such that:

• s_i ≠ s_j, and
• the ordered pair (s_i, s_j) appears exactly once among all pairs of indices (i, j) with 1 ≤ i < j ≤ n.

If no valid pair exists, output -1.

Note: All formulas are given in LaTeX format. For example, the condition j - i \ge 2 is written in LaTeX as shown.

inputFormat

The first line contains an integer n (n ≥ 1), the length of the gene sequence.

The second line contains n positive integers separated by spaces, representing the gene sequence s₁, s₂, ..., s_n.

outputFormat

Output a single integer, the minimum possible length of the subsequence s_i ... s_j (i.e. j - i + 1) that can be processed under the constraints. If no valid pair exists, output -1.

sample

5
1 2 3 4 5

3