Binary L2 mass-loss fraction for thin accretion disk

[pmocz]

This PR begins to implement some useful features for modeling thin accretion disks and mass transfer/loss in binary systems.

Here, we implement an internal function (made callable in run_stars_extras as binary_L2_mass_loss_fraction()) to

• calculate the (outer) L2 mass-loss fraction according to Lu et al. (2023) "On rapid binary mass transfer -I. Physical model"

and adds unit tests

[pmocz]

Currently, this function calculates the fL2_{outer} mass loss using Kramers opacity law. It looks like this (e.g., for q=0.5, M_2=10 Msun, X=0.7, Z=0.02)

[pmocz]

verified against: https://github.com/wenbinlu/L2massloss

[warrickball]

I've suggested powN(x) instead of x**N for N > 2, following our code style for exponents, and a bit more constant re-use (even for such details as ⅔).

Also note that the published version of the reference is Lu et al. (2023), not (2022).

[pmocz]

Thanks for all the fixes to help with bit-for-bit reproduction across compilers, @warrickball !