`newton!` and `find_closed_orbit` mishandling solution

[eiad-hamwi]

It looks like find_closed_orbit and newton! are mishandling the solution after correctly converging:

Specifically, newton!( f!, y, x, ...) places the final solution into x, and improperly labels y as the solution vector. In fact y is just the residual.

Then, find_closed_orbit(ring; v0=zeros(6)) basically throws away v0 == x and uses v == y as the solution.

The fix is to simply replace x <==> y in the order of arguments of newton! and then also update its return statements to give (; u=x, ...)

[mattsignorelli]

Fixed in #15