Convert $$y(x)=\lambda \int_{-1}^1 (1-|x-t|)y(t)dt$$ to differential equation with boundary conditions.
I know how to deal with this kind of problems but not when the kernel is a modulus function. I started solving it in the typical manner:
$$y'(x)=\lambda \int_{-1}^1 \frac{t-x}{|t-x|} y(t) dt.$$ It sort of rings the bell that $\frac{t-x}{|t-x|}$ could be $-1$ or $1$ depending on the values $x$ and $t$ on $[-1,1]$. Is there a way to determine those values? If I am right, it would be easy to solve this equation.
I also found another similar question (Transforming the integral equation $u(x) + \frac{\lambda}{2}\int_{0}^{1}|x - s|u(s)ds = ax + b$ into its equivalent differential equation) so I tried one of the methods in the answers:
$$\begin{align} \int_{-1}^1 (1-|x-t|) y(t) dt = \int_{-1}^{-x} (1-|x-t|) y(t) dt + \int_{-x}^0 (1-|x-t|) y(t) dt \\\\+ \int_{0}^x (1-|x-t|) y(t) dt + \int_{x}^1 (1-|x-t|) y(t) dt \end{align}$$
Then I'm really not sure what happens with the signs:
$$\begin{align} \int_{-1}^1 (1-|x-t|) y(t) dt = \int_{-1}^{-x} (1+x-t) y(t) dt + \int_{-x}^0 (1-x+t) y(t) dt \\\\+ \int_{0}^x (1-x+t) y(t) dt + \int_{x}^1 (1+x-t) y(t) dt \end{align}$$
Then by Leibnitz's Rule it should (?) follow that $$\frac{d^2}{dx^2} \int_{-1}^1 (1-|x-t|)u(t)dt=-4u(x)?$$ I feel like I just messed up with the signs...