[WIP] disable integrator.do_species_clip by default

[zingale]

it seems to cause more problems then it fixes

[zingale]

This (disabling clipping) seems to work well for a lot of problems, but for iso7 we get FPEs. I think that this is because there are some very negative mass fractions that are predicted in the nonlinear solver loop in vode_dvnlsd.H and when we call the RHS, the EOS call sees a negative $\bar{Z}$.

A robust solution might be to check the mass fractions inside that loop (around line 111) and then,
if they are negative, exit the function in the same manner we do if the matrix is singular. This will
force VODE to then try a smaller timestep.

[zingale]

The iso7 FPEs are fixed if we use the option in #1921

[zingale]

going to merge this into #1921, since I think both changes need to be done together.