Edge compression: extraneous stuff

[yakra]

• [x] Extraneous collapse conditions https://github.com/yakra/DataProcessing/blob/f86fadc9479d6ad91852c6c38d35f641b70ec6ef/siteupdate/cplusplus/classes/GraphGeneration/HighwayGraph.cpp#L105-L106

  No need to check for reversed order of incident edges because they're always in order of their construction. Nota Bene: This may change in the future:

  // replace edge references at our endpoints with ourself
  if (fmt_mask & collapsed)
  {   (vertex1->c_front == edge1 ? vertex1->c_front : vertex1->c_back) = this;
      (vertex2->c_front == edge2 ? vertex2->c_front : vertex2->c_back) = this;
  }
  if (fmt_mask & traveled)
  {   (vertex1->t_front == edge1 ? vertex1->t_front : vertex1->t_back) = this;
      (vertex2->t_front == edge2 ? vertex2->t_front : vertex2->t_back) = this;
  }

  ...thereby replacing an earlier vertex with a later vertex. Don't be tempted to think a collapsed compression will always have a corresponding traveled compression (whether combined or separate. Collapsed-only compression can cause reversals of the collapsed edge order only. Once that changes, either...

  • Restore the equivalent. Do some counts, DevelInfo: combined compress forward @ A370/AH30@+X421910 style,
  • or https://github.com/yakra/DataProcessing/issues/95#issuecomment-778836635

• [x] Extraneous pointer chasing https://github.com/yakra/DataProcessing/blob/f86fadc9479d6ad91852c6c38d35f641b70ec6ef/siteupdate/cplusplus/classes/GraphGeneration/HighwayGraph.cpp#L76

  for (HGVertex& v : vertices) v.foo = bar etc. Maybe hold off on this a bit? Conflicts with some of the alternatives for preventing segment name mismatch errors.

• [x] Making this unnecessary https://github.com/yakra/DataProcessing/blob/f86fadc9479d6ad91852c6c38d35f641b70ec6ef/siteupdate/cplusplus/classes/GraphGeneration/HighwayGraph.cpp#L51-L52