I am working through some distribution theory notes, and was specifically working on this example Derivative of principal value distribution $1/x^2$ is equal to finite part distribution $-1/x^2$?.
You can reformulate the parte finie as $\lim\limits_{\epsilon \to 0} \int\limits_{|x| \geq \epsilon} \frac{\phi(x) - \phi(0)}{x^2} dx$.
I'm fine with everything except the very first statement, as I do not understand why this integral is finite for $\phi \in \mathscr{D}(\mathbb{R})$ (the class of infinitely differentiable functions with compact support). I attempted to prove it, but of course because of the fact that there is no control of $\frac{1}{x^2}$ in a neighborhood of $0$ we can't control the integral at that particular singularity in the usual way (bounding the quantity $|\phi(x) - \phi(0)|$ by something small). I attempted to consider a similar proof to that of proving the Riemann integral of $\frac{\sin(x)}{x}$ converges, but that doesn't work either, which got me thinking of a counterexample. We can most certainly find a $\phi \in \mathscr{D}(\mathbb{R})$ such that $\phi(x)|_{[0,1]} = \sin(x)$. Then $\int_0^1 \frac{\phi(x)-\phi(0)}{x^2} dx = \int_0^1 \frac{\sin(x)}{x^2}$, but this is infinite. Throw some absolute values on, bound above by the integral over all of $\mathbb{R}$, it seems to indicate that the parte finie should be infinite. I feel like I'm either missing something to do with the Cauchy integral theorem (although with a generic function like $\phi \in \mathscr{D}(\mathbb{R})$ I'm not entirely sure how to use it) or something else entirely. Any help would be appreciated. Thanks!