Wrong tester files for BarycentricInterpolation branch
The problem lies in the tester files where we have not provided -P arguments. These are responsible for creating different PointsType than PolyEvenlySpaced (where the function evaluations are provided). In the absence of this argument, the demos were taking exactly the same set of points for evaluation and interpolation. Therefore, the check for (z[i] == coords) was true for all z[i] and instead of interpolated value, we just received the input function value.