Co-authored-by: Max Horn <[email protected]>

Before, it might have happened that the command `R * ZZRingElem(2)` fails while `R * big(2)` succeeds.

Max: "the advantage is that if some ring type wants to implement say ideal(R::MyRing, n::Int) in a special optimized fashion that avoids constructing R(n), then R*2 would benefit from that"

Co-authored-by: Max Horn <[email protected]>

Previously, there was a possibility of an infinite recursion if the base functions were not implemented. Also, writing `R(x)` when `x` is a vector is problematic, it might have unexpected results.

This is already allowed for Hecke ideals.
For consistency, we allow this already in AbstractAlgebra.