Hi, I don't know if this is a bug: Does eliminate_metric supposed to move the metric tensor inside partial derivative freely?

I have the following code:

```
{x1,x2,x3,x4}::Coordinate;
{k,l,m,n,p,q,r,s,t}::Indices(position=fixed,values=[x1,x2,x3,x4]);
g{m n}::Metric;
g^{m n}::InverseMetric;
g{m}^{n}::KroneckerDelta;
g^{m}{n}::KroneckerDelta;
{g{m n},g^{m n}}::Depends(x1,x2,x3,x4);
\nabla{#}::Derivative.
\partial{#}::PartialDerivative.
test := Q^{s}_{r m n} -> g^{s k} \partial{r m}{g{k n}};
eliminate_metric();
```

The output I get is that the inverse-metric goes inside the derivative and contracts with the metric to form the Kronecker-delta. Is this supposed to happen? Am I supposed to limit the method eliminate_metric to avoid partial-derivatives?