Wrong results when using singular functions

[enriqueartal]

In sage (version 8.1, but 8.0 has the same issue) I have the following code:
R.<x,y>=PolynomialRing?(QQ,order='neglex')
f=(y2+x3)(y2+x3+x2y)
singular.lib("all.lib")
f.singular().bernstein()

I guess this is equivalent to the following singular code
LIB "all.lib";
ring r=0,(x,y),ds;
poly f=(y2+x3)*(y2+x3+x2y)
bernstein(f)

The results differ: -11/7 is a root in the first case, while -4/7 in the second (and I think there are theoretical results that ensure the second one is correct); moreover in both cases, when using bfct -4/7 is the root (bfct uses another algorithm to compute the Bernstein-Sato polynomial). The Singular computations are done using both an external Singular installation and the Sage one.

In other cases I had communication problems between singular and sage leading to output errors, but in this case it produces (a non evident) mathematical wrong result.

Upstream: Reported upstream. No feedback yet.

Component: algebraic geometry

Keywords: singular

Issue created by migration from https://trac.sagemath.org/ticket/24377

[enriqueartal]

comment:1

After discussion with upstream, it is related to the dependence on orderings of the rings, so I guess it can be closed.