by JoelKatz » Thu Apr 13, 2017 12:56 am
Are you asking about the old implementation of getNodeFat? It was an ugly hack to keep a node from getting added twice. We had to add the node late in the loop because we didn't know at that point if the loop would repeat or not. But if the loop repeated, the node would wind up getting added a second time. So we set a "skipNode" flag indicating that we should skip adding that node, because we already added it. The logic was later changed to avoid this hackiness.