"toField" prevents getting the right answer for the kernel of a ring map

Nobody has claimed this yet.

```m2
Macaulay2, version 1.16.0.2
with packages: ConwayPolynomials, Elimination, IntegralClosure, InverseSystems, LLLBases, MinimalPrimes, PrimaryDecomposition, ReesAlgebra, Saturation, TangentCone, Truncations

i1 : A=toField(QQ[w,DegreeRank=>0]/(w^3-1));

i2 : R=A[x,y,z,Degrees=>{4,4,4}];

i3 : S=A[s,t];

i4 : qmap = map(S,R,{t^4-w*s^4,s*t*(t^2+s^2),s*t*(t^2-s^2)});

o4 : RingMap S <--- R

i5 : I = kernel qmap

o5 = ideal (0, 0, 0, 0, 0, 0)

o5 : Ideal of R
```

Since R has more variables than S, the kernel can't be 0.

Run the reported Macaulay2 session and inspect the behavior of toField and kernel on the shown ring map. Trace the kernel computation for this example and identify why it returns the zero ideal; done means the example reports the correct nonzero kernel.

Written by the indexing model from the issue text.