[WIP] Adapt the IActsGeoSvc interface to fit better to Acts downstream needs

[tmadlener]

BEGINRELEASENOTES

• Make the IActsGeoSvc return the tracking geometry as a shared_ptr to better align with downstream usage
• Add a magneticField method to the IActsGeoSvc and return a constant B field for now.

ENDRELEASENOTES

@wdconinc Do I remember correctly that you are still using this and this will probably be breaking changes?

[wdconinc]

Do I remember correctly that you are still using this and this will probably be breaking changes?

We're only using this in legacy mode. We have the permanent desire to get back to improving our gaudi support since all our algorithms are written in a framework-agnostic and this is exercised through a gaudi pipeline, but it's obviously a lower priority.

As long as there's a release before major changes take effect, we can pin on that. It may be there already... Or we can pin on a commit too.

[tmadlener]

Sounds good. We have https://github.com/key4hep/k4ActsTracking/releases/tag/v00-01 from end of October. No behavior changing changes have landed since, so IIUC we could break things for the next tag, right?

[wdconinc]

so IIUC we could break things for the next tag, right?

Break all you want :-)

[tmadlener]

Suggested change

	using CellIDSurfaceMap = std::unordered_map<uint64_t, const Acts::Surface*>;

[tmadlener]

Closing this in favor of #36, where all of these changes have a better motivation to be given that the original implementation of the ActsGeoSvc was not really used by the functionality added in #25