If you've got access to the rational root theorem, that's immediate
actually, though
maybe not immediate
a brute-force approach would be to say that the part of the quadratic formula that leads to an irrational answer is the square-root part, i.e. $b^2-4ac$
so you want to show that this is never a perfect square.
Okay, let me see if I did the next problem right: if I have the function $g(\theta) = \frac{\tan \theta}{\theta^2 - \frac{\pi}{9}}$ then I have the domain $\{\theta\in\mathbb{R}|\theta \neq \sqrt{\frac{\pi}{9}}\}$, right?
@JessyCat Given any divisor $d\mid n$ (let's say $n=dk$), the subgroup of $\Bbb Z/n\Bbb Z$ generated by $d$ is given by $\{0,d,2d,\cdots,(k-1)d\}$. This is cyclic of order $k$. There are a set of coset representatives given by $\{0,1,2,\cdots,d-1\}$ (all integers mod $n$ of course). Given two representatives, to compute the sum in the quotient group you just add the representatives and discard excess multiples of $d$. In this way, the quotient group is cyclic of order $d$.