Use interface types for TrackerHit

[tmadlener]

BEGINRELEASENOTES

• Make the edm4hep::TrackerHit an interface type and use that in the edm4hep::Track.
• Rename the current edm4hep::TrackerHit to edm4hep::TrackerHit3D
• This is a breaking change. However, we currently do not plan to implement schema evolution for this

ENDRELEASENOTES

• Needs Add support for interface types AIDASoft/podio#516
  Breaking changes downstream:
• Make upcoming renaming of TrackerHit to TrackerHit3D transparent k4EDM4hep2LcioConv#45
• Make it possible to transparently migrate to TrackerHit interface k4FWCore#174
• Make the renaming of TrackerHit -> TrackerHit3D transparent AIDASoft/DD4hep#1219 (DDDigi also uses the TrackerHit and needs to be updated accordingly.)
• Switch to TrackerHit3D for introduction of TrackerHit interface k4MarlinWrapper#168
• Switch to TrackerHit3D transparently k4SimDelphes#113
• Use TrackerHit3D once it becomes available HEP-FCC/k4SimGeant4#62
• Prepare for upcoming renaming and interface introduction HEP-FCC/FCCAnalyses#345
• CEPCSW breaks (CEPCSW out of the Key4hep nightlies cepc/CEPCSW#263)
• Update documentation (add an interfaces section?)

This is one of a few rather breaking changes that we plan to do in the near future towards an EDM4hep version 1.0.

Fixes #218

[jmcarcell]

I'm not sure how to fix the CI for this one. So it fails because it picks up the FCCAnalyses dictionaries and headers and there something wants TrackerHitData.h which doesn't exist anymore with this PR. Why ROOT does this when it's not needed I don't know. Locally the tests pass for me without FCCAnalyses so one option would be to remove FCCAnalyses in the environment for the tests as they should be only related to EDM4hep. However this is useful information since people using EDM4hep in Key4hep will have FCCAnalyses in the environment, so we can be aware if something is failing. I think I'll just modify the environment and add a test to the tests that run post-build so that we would be aware if there were issues like this one.

[tmadlener]

The problem here is that LD_LIBRARY_PATH is still in the test environment. FCCAnalyses from the nightlies is still built against the older EDM4hep, so it still has a TrackerHitData. ROOT picks up things pretty indiscriminantely from there. This would be solvable in EDM4hep probably by constructing the test environment to contain exactly what it needs, because there is no FCCAnalyses in the LCG stacks. LCG stack environments are pretty impossible to deal with in these cases because there is almost always something that pulls in <lcg-view-path>/lib[64].