Define
$$F(t,x)=\frac{\partial}{\partial t}\bigg(\text{sgn}(t)\,\delta(t^2-x^2)\bigg),$$
where $\text{sgn}(x)$ is the sign function, and $\delta(x)$ is the Dirac $\delta$-function.
Does $F(0,x)$ make sense, and what is it?
Define
$$F(t,x)=\frac{\partial}{\partial t}\bigg(\text{sgn}(t)\,\delta(t^2-x^2)\bigg),$$
where $\text{sgn}(x)$ is the sign function, and $\delta(x)$ is the Dirac $\delta$-function.
Does $F(0,x)$ make sense, and what is it?
Heuristic approach: Before differentiation wrt. $t$, one may rewrite
$$ {\rm sgn}(t) \delta(t^2-x^2)~=~ \frac{1}{2x} \left\{ \delta(t-x) - \delta(t+x)\right\}. \tag{1} $$
Differentiation wrt. $t$ yields
$$F(t,x)~:=~\frac{\partial}{\partial t}\left\{{\rm sgn}(t) \delta(t^2-x^2) \right\} ~=~ \frac{1}{2x} \left\{ \delta^{\prime}(t-x) - \delta^{\prime}(t+x)\right\} .\tag{2}$$
Then $$F(t=0,x)~=~-\frac{1}{x} \delta^{\prime}(x).\tag{3} $$
This can be made mathematically precise by using test functions.