lp.rhs seems incorrect

I'm working on a Gomory cut generator that uses CyLP. I was assuming that lp.rhs contained the value of B^-1 b and this seems to be the case after initially solving a giving linear program. However, if I add an inequality (cut) and then resolve, lp.rhs is no longer correct. This can be checked easily by comparing the value of the solution that CyLP returns with the value of the right-hand side. The right-hand side values for the rows in which the original variables are basic should match the solution values and they don't. To reproduce, get CuPPy and run it out of the box.

https://github.com/tkralphs/CuPPy

You will see that in the second iteration, lp.rhs and sol don't match.