Notes:

• We should probably also rework how we store the input stop data and how we figure out when we need to add stops

List of constraints and why we did it that way:

• TimeData values: this is the place where we edit the most the stop data, where we make them longer or shorter. This structure is a nice place to store this data.
• InfraExplorerWithEnvelope:
  • We also need the stop offsets
  • Stops need to be part of the conflict detection inputs. This class is supposed to generate the conflict detection inputs.
  • We need to build an EnvelopeStopWrapper (or something with a similar behavior)
• InfraExplorer:
  • We need to list stops in PathFragments
  • We need the path start and end

Some partial conclusions:

• We need to figure out if/when we stop as we explore the path, not when we simulate that part
  • We also need to figure out if the stop is on a closed signal
  • (We can just always stop on closed signals)
• We could model it as a stop of infinite duration and we would edit it later on. But we need to carefully check that the spacing resource generator doesn't cache the wrong stuff.
  • We may just not cache ressource use if there's an infinite-length stop
  • This matches the use case of pendingRequirements, we can see if it can handle it (with or without change)

TODO?

• Goal: keep stop durations in TimeData, keep stop offsets in InfraExplorer. Define stops as we extend the InfraExplorer's lookahead.
• To remove the InfraExplorerWithEnvelope part:
  • We can move stop offsets and move them (to InfraExplorer maybe?)
  • Stop as inputs to the conflict detection: have to be handled by the InfraExplorer (part of IncrementalPath)
  • We have all the info we need to build the EnvelopeStopWrapper, using TimeData
• InfraExplorer
  • Improve how we find stop locations when extending infra explorers (TODO, maybe create a new class to handle it? It would handle closed signal decision)