I would expect that when one evaluates a polynomial, only the coercion
properties of the *base ring* of the polynomial ring relative to the
ring of definition of the evaluation point are important, but the
example below gave me an unexpected negative answer.
The fact that R is has an automatic
Nils,
See http://trac.sagemath.org/sage_trac/ticket/8502. I had found that,
depending on the base ring, the result of evaluating all the variables
was sometimes in the coefficient ring and sometimes a (constant)
polynomial. So I fixed it -- I think. But the result is clearly not
perfect.
John