Fix the result of innerMaterial() and outerMaterial() depending on if the other is called first

[jmcarcell]

BEGINRELEASENOTES

• Fix the result of innerMaterial() and outerMaterial() depending on if the other is called first
• Add a test that fails without this fix and passes with the fix

ENDRELEASENOTES

Fix #1413.

I don't know exactly why this is happening for a few surfaces. What happens is that when getting the next boundary after innerMaterial() has been called is not the same as if it wasn't called (in this call

DD4hep/DDRec/src/MaterialManager.cpp

Line 77 in 24fcc3e

TGeoNode * node2 = _tgeoMgr->FindNextBoundaryAndStep( 500, 1) ;

). Making a backup of the navigator and restoring it solves this problem.

[andresailer]

Thanks! That was a nice big piece of debugging!

[andresailer]

@MarkusFrankATcernch

We have a failing seed for the box of straws test here

  EventSeeder1     INFO  At beginEvent: eventID=0, runID=0 initialSeed=1960713805, newSeed=3837709546

[github-actions]

Test Results

   16 files     16 suites   6h 5m 51s ⏱️
  371 tests   370 ✅ 0 💤 1 ❌
2 895 runs  2 894 ✅ 0 💤 1 ❌

For more details on these failures, see this check.

Results for commit 47a6e2b.