ha. maybe say that the integral gives the same value if the upper limit was (pi/2 - h) as h -> 0, h > 0 or something? i dono
edit: or break the initial integral into two integrals, one of which contains the 0 limit, and the other has the pi/2 limit. so then you can divide by cos^2x for the first one, and sin^2x for the second so you avoid both cot0 and tan(pi/2) ?