The following questions is from Real and Complex Analysis editted by Walter Rudin. It is the 5th exercise in Chapter 9 (the one about Fourier Transform).
Let $m$ be the Lesbegue Measure on the real line and define $\int_{-\infty}^{\infty}f(x)dm(x) = \frac{1}{\sqrt{2\pi}}\int_{\infty}^{\infty}f(x)d(x)$Let $S$ be all functions $f: \mathbb{R} \rightarrow \mathbb{C}$ that are infinitely differentiable and for any $m, n \in \mathbb{N}$, there exists a real number $A_{mn}(f)$ such that $sup|x^nf^{(m)}(x)| \leq A_{mn}(f)$. Now prove that the fourier transform mapping maps $S$ onto $S$.
According to Theorem 9.2, point (2) inside the book, when $f \in L^1$, $(\hat{f})^{(1)}(t)$ = $-i \int^{\infty}_{-\infty} xf(x)e^{-ixt}dm(x)$. However, in this question, I do not know whether or not $f$ is in $L^1$. If it is, how to find an upper bound when taking integration along the real line?
The point of $A_{mn}(f)$ is control of tail behavior: these functions decrease faster than every polynomial. So we should use this condition to estimate $\int |f|$ near infinity:
$$ \int_1^\infty |f|\ dm = \int_1^\infty |x^{-2}||x^2f|\ dm \leq \sup |x^2f| \int_1^\infty |x^{-2}|\ dm \leq A_{21}\int_1^\infty x^{-2}\ dm < \infty$$
Use other means to control $\int|f|$ on compact intervals and you're set.