Tim Adye

Sorry, I may have misunderstood the way this is intended to work. I guess you want `addAmbiguityResolution` to set things up so the subsequent step (`addVertexFitting`) does the right thing...

Currently this has the default `twoWay = true` (in `TrackFindingAlgorithm.hpp` and `reconstruction.py`). Should we make it `false` to keep the old behaviour, or make this the new default?

The changes look OK to me. Unfortunately the results aren't so good running on ITk with ttbar PU200. 1. It crashes after 81 events with error: ``` /home/ppd/adye/acts/build/src/Core/include/Acts/Utilities/VectorHelpers.hpp:142: std::array Acts::VectorHelpers::evaluateTrigonomics(const...

Just checked the latest version with the ITk, including `restore first track` patch. The number of `ERROR` messages are now similar between runs with `twoWay=true` and `false`. I still get...

I ran the ITk ttbar+PU200 tests again (this is standalone ACTS). This still crashed (`twoWay=true`) or went into a loop (`twoWay=false`) within 100 events. As suggested before, this is probably...

I have removed the workaround for this issue from #1467 and confirmed that the errors still occur with the latest `main` (ITk ttbar+PU200): ``` 19:34:43 Propagator ERROR Step failed with...

We get 3518 interpolation errors per ttbar+PU200 event, so it's not difficult to reproduce 😄. Most errors print the initial parameters, eg. ``` 15:23:18 Propagator ERROR Step failed with MagneticFieldError:1:...

yes, default `full_chain_itk.py` just enabling `ttbar_pu200=True`. One event is enough! You even see it with ttbar `npileup=1` - but with a manageable number of errors.

This bug is largely fixed by #2299, and completely by #2300. There are no remaining `interpolation out of bounds` messages from `full_chain_itk.py` with the `ttbar_pu200` option. Those fixes each improve...

> how do we move forward with this @paulgessinger @timadye ? Good question. On this PR, we can take just the first changed line, removing the incomplete but unneeded initialisation....