The performance of this is N ** 2 + NM (where N is the size of s2 and M is the size of s1) since:

• it loops through all the characters of s2 (N)
• for each of those caracters it counts the time the character 'x' appears. For that it has to loop through all s2 (N) and s1 (M)

That's far from ideal, which should be N + M

uniq method isn't needed

It's nice, I had similar solution, but even now I cannot pass performance tests with it.

Fails the "performance matters" test.

cool