The Vigenere Cipher is a well known cipher in cryptography. The encryption of a string $s$, with the help of a key-string $t$, both having the same lengths, is done by the following way:

Let $e$ be the final string that will be obtained by the encryption of string $s$ using the key $t$.
The string $e$ will have the same length as the previous $2$ strings, and will respect the following property:

where $\textrm{ord}(c)$ is the ordinal number of the character $c$ in the english alphabet ($\textrm{ord}(\texttt{`a'})=0$, $\textrm{ord}(\texttt{`z'})=25$ and so on), and $\textrm{chr}(i)$ is the function that returns the character with the ordinal number $i$ in the english alphabet.

Briefly, the encryption process can be described as an iterative addition, character by character, modulo 26, of the string $t$ to the string $s$.

Let's define a modification over the Vigenere Cipher, which we will refer to as: Cyclic Vigenere, the following way:

The encryption of string $s$ with key-string $t$ is possible only when the length of string $t$ divides the length of string $s$. The result of the encryption is string $e$, defined as:

where $|t|$ denotes the length of the $t$ string, $i \mod |t|$ is the remainder of $i$ by $|t|$, and the indexing of the strings starts at 0.

Briefly, the encryption process can be described as an iterative addition, character by character, modulo 26, of the repetition of the string $t$ to the string $s$.

In this task, $K$ key-strings are given, and $N$ strings that need to be encrypted. For any string $s_{i}$ from the $N$ strings, print the index of the key-string $t_{j}$ that generates the minimal possible lexicographic encryption $e$ of the string $s_{i}$. If there are multiple answers, the one with the smallest index must be printed. If there is no answer, print $-1$.

Input data

The first line of input contains $N$ and $K$.

The next $K$ lines contain the key-strings $t_0, t_1, \ldots, t_{K-1}$.

The next $N$ lines contain the strings $s_0, s_1, \ldots, s_{N-1}$.

Output data

Output $N$ lines. The $i$-th of them must contain the index of the key-string $t_{j}$ that generates the minimal possible lexicographic encryption $e$ of the string $s_{i}$. The indexes are considered to start from $1$.

If there are multiple answers for any of the strings, print the minimal possible index of a key-string $t_{j}$.

If there is no such key-string, print $-1$.

Constraints and Clarifications

• $1 \le N \le 2 \cdot 10^5$.
• $1 \le K \le 2 \cdot 10^5$.
• The total lengths of all strings and key-strings is at most $4 \cdot 10^5$.
• Each string contains only lowercase English letters.

Example

stdin

3 6
are
mere
ana
oare
cinestie
nimeni
anaana
plebi
nusestie

stdout

3
-1
4

Explanation

In the sample, for the first string anaana, the key-string are produces aeeaee, the key-string ana produces aaaaaa, and the key-string nimeni produces nvmeai, the lexicographically minimal encryption result is aaaaaa, which is produced by the third key-string, so the answer is $3$.

For the second string plebi, there is no suitable key-string, since it has length 5, which is not divisible by any of the lengths of the key-strings, thus the answer is $-1$.

For the third string nusestie, the key-string mere produces zyjiexzi, the key-string oare, produces bujigtzi, and the key-string cinestie produces pcfikmqi, the lexicographically minimal encryption result is bujigtzi, which is produced by the fourth key-string, so the answer is $4$.