Can either of the following integrals be expressed in explicit forms in terms of known constants? They can certainly be expressed in terms of special values of hypergeometric functions with many parameters, but I am interested in expressions in terms of some combination of simpler (multiple)-zeta values and L-values if such expressions exist.
Here I am interested in
$$I_2 = \int_{0}^{1/4} \frac{(\log(1-x))^2}{x \sqrt{1-4x}}dx,$$
$$J_2 = \int_{0}^{1/4} \frac{\log((1-x))^2}{(1-x) \sqrt{1-4x}}dx.$$
Some added clarification in light of a comment, this is the square of $\log(1-x)$, not the log of $(1-x)^2$ which simplifies. It turns out for comparison that
$$I_1 = \int_{0}^{1/4} \frac{(\log(1-x))}{x \sqrt{1-4x}}dx = - \frac{\pi^2}{18},$$
$$J_1 = \int_{0}^{1/4} \frac{\log((1-x))}{(1-x) \sqrt{1-4x}}dx = \frac{\pi \sqrt{3} \log 3}{9} - L(2,\chi_{-3})$$
where $L(2,\chi_{-3})$ is the Dirichlet L-value $\displaystyle{ \sum \frac{1}{(3n+1)^2} - \frac{1}{(3n+2)^2}}$.
Mathematica can "evaluate" the second integral (even as an indefinite integral) in terms of hypergeometric functions and so reduces the calculation in that case to ${}_{4}F_{3}(1/2,1/2,1/2,1/2;3/2,3/2,3/2;3/4)$, but I have no reason to expect either way whether this has a simpler form or not.
One reason to possibly imagine there might be an explicit answer is because the simpler versions $I_1$ and $J_1$ can be evaluated. One reason to imagine there might not be is that they are related to special values of Hypergeometric functions not at $z=1$ but at $z=3/4$ or $z=1/4$, and there seem to be fewer examples known of special values at this point.