HZ: avoid convergence if bisected

[timholy]

secant2 calls update, and update might switch to bisection in the U3 step (as named in the initial HZ publication). If we did bisect, the line-search "are we making enough progress?" check (step L2) is nonsensical (we might have shrunk multiple iterations of bisection, but that is not an indication that the secant model was "working").

Consequently, this reports back about whether bisection was engaged in update, and if so skip any kind of convergence assessment and do another iteration.

Fixes #173. Sadly there isn't a very portable test.