Support the decimal crate

[eugene-bulkin]

Is there any way we can support the d128 used in the decimal crate for use in matrices? e.g. I want to have a matrix of d128 values and do stuff with them.

Note: This would require impl-ing BaseNum and Cast<f64> for d128. The last one seems sketchy to me since that cast is not precise.

[sebcrozet]

We could optionally support this behind a feature. Implementors of the Cast trait are not bound to strong accuracy requirements. They are just expected to return the closest représentable value.
The main difficulty here is that BaseFloat requires num::Float to be implemented and it looks like d128 is still missing all the trigonometric functions.

[milibopp]

@eugene-bulkin in today's world like all traits that need to be implemented are outside of nalgebra. So if this is still a relevant issue, I would suggest to pursue the integration of decimal with alga and num-traits to make these operations work.

I don't think there is anything to be done here though. Therefore I am going to close this issue.