Make low degree solvers work with complex numbers.

[emsr]

Make low degree solvers work with complex numbers.
If they don't already.

[emsr]

The algorithms should actually pretty much just work.
Making things generic might be a trick.
I quadratic there's a test and a branch to prevent loss of precision - making that complex-aware might take a minute.