let g(x) be a sufficiently smooth function and k,r $\in \rm I\!R$
via partial integration ($\int uv' = [uv] - \int u'v$), I arrive at
$$I:\lim_{r\to\infty}\left( \left(\frac{g(2)}{ikr}\exp^{2ikr}-\frac{g(0)}{ikr}\right)-\left[ \int_0^2 \frac{1}{(ikr)}g'(x) \exp^{ikrx} dx\right]\right)$$
When I then choose g'(x) to be v' and the exponential term to be u, I get
$$II:\lim_{r\to\infty}\left( \left(\frac{g(2)}{irk}\exp^{2ikr}-\frac{g(0)}{ikr}\right)-\left[ \left[g(x) \frac{1}{ikr} \exp^{ikrx} \right]^2_0 -\int_0^2 g(x) \exp^{ikrx} dx\right]\right)$$
The terms in front of the integral cancel each other and I arrive back at the start. When I choose u & v' the other way arround I'll get
$$III:\lim_{r\to\infty}\left( \left[g(x) \frac{1}{ikr} \exp^{ikrx} \right]^2_0-\left[ \left[g'(x) \frac{1}{(ikr)^2} \exp^{ikrx} \right]^2_0 -\int_0^2 g''(x) \frac{1}{(ikr)^2} \exp^{ikrx} dx\right]\right)$$
I don't see how this approach could bring me further, since the integral didn't simplify. I haven't used the limit $\lim_{r\to\infty}$ yet, so I guess this should be rather important. But on the other hand side, I also don't see how the limit would help, since the exponential function will always go to infinity faster than the fraction.
I missed that the exponential function is oscilating, so it will go to zero for r going to infinity. This clears removes the terms in front of the integral in III. For the integral $\int_0^2 g''(x) \frac{1}{(ikr)^2} \exp^{ikrx} dx$, can I argue that it will also go to zero, since the integral of $exp^{ikr}$ will always be oscilating an the prefactor $\frac{1}{(ikr)}$ overpowers the rest?