Meld assumes Bianchi identity for derivative of Riemann

[kpeeters]

{a,b,c,d,e,f}::Indices(position=independent);
\nabla{#}::Derivative;
R_{a b c d}::RiemannTensor;
bianchi:= \nabla_{a}{ R_{b c d e} };
asym(bianchi, $_{a}, _{b}, _{c}$);
meld(bianchi);

produces zero, even though there may not be any formal connection between this derivative and the metric that builds the Riemann tensor.

Moreover, if

\nabla_{a}{ R_{b c d e} }::SatisfiesBianchi;

is explicitly declared, then meld does not produce zero.

It would probably be best to formalise connections between RiemannTensor, SatisfiesBianchi, Derivative and Metric so this can actually be checked properly.