I can't quite understand the scope of Depends(). If I run this:

```
{a, b, c, d, e, f#}::Indices(full, position=independent).
\nabla{#}::Derivative.
\partial{#}::PartialDerivative.
A{#}::Depends(\partial{#}).
A{#}::Depends(\nabla{#}).
ex:= \partial_{ a }{\nabla_{ b }{ A_{ c d }}} -> \nabla_{ a }{ \nabla_{ b } {A_{ c d }}} + \Gamma^{ e }_{ a b } \nabla_{e}{A_{ c d }} + \Gamma^{ e }_{ a c } \nabla_{ b }{ A_{ e d }}
+ \Gamma^{ e }_{ a d } \nabla_{ b }{A_{ c e }};
unwrap(_);
```

I see that ex doesn't contain the double nabla on the rhs after evaluation. Moreover, unwrap() reveals that the lhs of ex is evaluated to zero.

Why is that? How can this be treated correctly?