Isolation boundaries currently feel a bit linear, and while they do avoid infinite loops (where A goes to B goes to A etc.) it currently does not take order in the node graph into account.
Ideally, if only the "end" of an isolation boundary has changed, we only need to recalculate that one pool rather than the entire boundary set.
Additionally, if the "end" of an isolation boundary has an id value that places it in the middle of the boundary set, we don't want to recalculate everything before/after etc.
Isolation boundaries currently feel a bit linear, and while they do avoid infinite loops (where A goes to B goes to A etc.) it currently does not take order in the node graph into account.
Ideally, if only the "end" of an isolation boundary has changed, we only need to recalculate that one pool rather than the entire boundary set.
Additionally, if the "end" of an isolation boundary has an
idvalue that places it in the middle of the boundary set, we don't want to recalculate everything before/after etc.