monitor condition number of Polytope

[johnyf]

To make the user aware of how bad ill-conditioning is in their problem, a quick search leads, e.g., to this paper.

[slivingston]

Can you make this issue more specific? E.g., are there specific motivating examples that new features should address? E.g., as a feature, would this be a package-wide flag that is fixed at time of importing polytope that would cause functions to fail in the midst of a computation if intermediate values are deemed to be badly ill-conditioned?