First of all, please post your questions using mathjax. Scanned images are unacceptable and may get your question closed before getting an answer.
Now, to your problem. The correct answer is yours. However, both you and "they" have computed the frequency response of the filter incorrectly. In particular, you have both ignored the periodicity of the discrete-time Fourier transform (DTFT). Your "solution" implies that the DTFT of $e^{i\pi n}$ is $\delta(\omega - \pi)$. This is not correct. It actually holds
$$
\mathcal{F}(e^{i \pi n})=\sum_{k=-\infty}^{\infty}\delta(\omega - \pi - 2\pi k).
$$
Using this formula you can now see that the correct frequency response is as shown in the figure (the dashed lines delineate the period $\pi<\omega<\pi$; the Fourier transform is periodic and extends over all $\omega \in \mathbb{R}$).

The other solution is based on expressing $(-1)^n$ as $\cos(\pi n)=\frac{1}{2} e^{i \pi n} + \frac{1}{2} e^{-i \pi n}$. I will leave it to you to verify that the DTFT of $\cos({\pi n})$ is also as shown above.