Implement handling of nonlin Real division

[BritikovKI]

Currently #790 does not allow for the support of Real nonlinear division (with variables in the divisor).
We will need to implement this feature at some point if we would want to extend to nonlin solving

[blishko]

My argument would be that we actually do not need general division, because handling division by zero is extremely messy.
Support for polynomial arithmetic without division would be completely fine, in my opinion.