You are given n tests numbered from 1, 2, 3.... N. Each test passes individually, but there are two tests which when ran together fails. You can call a function testSuite() and pass any number of test as you want.

For example :
Given N = 5,
1, 2, 3, 4, 5

Lets say 1 and 4 are failing tests.

If you ran testSuite([1, 2, 3, 4, 5]) -> fail
If you ran testSuite([1, 2, 3]) -> pass
If you ran testSuite([1, 2, 4]) -> fail
Likewise, testSuite([1, 4]) -> fail

Can you write a algorithm (better than O(N^2) ofcourse) to find the failing tests?