Track Propagation Ends with Change in Volume ID

[mlarson02]

• Expected that ACTS would propagate in both directions (with propagate backward or propagate forward enabled), however, found that propagation in the opposite direction as configured ends when the propagation leaves the volume ID that the seed started in.
• e.g., from testing with a 50 GeV, eta = 0 muon gun, we see that when seeding in the last VXD layer, 1st 2 IT layers and using propagate forward, at least 4 VXD hits will be picked up (consistent with the number of expected hits in the same volume ID as the last VXD seeding layer), whereas when seeding in 3 IT layers only and propagating forward, no VXD hits will be picked up. Further details are available in this talk given in this MAIA meeting on slides 9 - 18.

• a fix for this exists beginning in ACTS version 37.0, allowing configuration of which volume IDs seed propagation can continue in. Therefore a new software release must either be built using a newer version of ACTS, or by integrating these changes into a new 32.1.0 ACTS build and generating a new software release with that.

[tmadlener]

I suppose most of this has already been discussed at the meeting last week. Nevertheless adding a brief comment here.

Given that the tracking is currently being moved to run in Gaudi rather then on Marlin in the efforts to upstream this to https://github.com/key4hep/k4ActsTracking, Once the Gaudi port has been validated with Acts v32, the next step will be to go to the latest Acts version (or at least close to that). I suppose that should solve this issue at least on a technical level.