Current

When I want to know if a puzzle is valid or not, I will check that all of the nodes are complete, and then that there is a single loop that sees all of the nodes. This means I need to walk from coord to coord along edges that exist, keeping track of the coords I see along the way.

Future

As we add edges, we should keep track of how they connect to each other, and what nodes they see along their path. Then, when we add more edges, we can just walk those and see if they create loops or if they see more Nodes.

type segmentCap struct{
  coord model.NodeCoord
  edge model.EdgePair
}

type pathSegment struct{
  start segmentCap
  end segmentCap

  seenNodes []model.Node
}

then on the Puzzle, we could keep a []pathSegment. And once we've runQueue on performUpdates, we can look at all of the coords we have in all of the segmentCaps, and only need to walk out from there. We could do this on every performUpdates, or only once we've completed all nodes. There is benefit both ways.