Iteration count off by one causing early false failure

[tekwen]

While working with a parcel fabric boundary network it was noticed that there appears to be an indexing error related to when convergence tolerance is reached, resulting in an adjustment that reports failure when in fact it has converged.

In speaking with @tbrinkman , it seems to be an issue on the Dynadjst side rather than in the Esri code.

Here's the experience:
In this case, I'd set the max iterations to 3. Run the adjustment and get an apparent convergence as far as the numbers would indicate, and yet a report that there is not a convergence:

Set the max iterations to 4, and get the exact same number of iterations/results, and success after the 3rd iteration, with no 4th iteration: