TreeBot shouldn't throw an UnsupportedOperationException if a child of a branch is null, in trees of graph theory, not having a right or left subtree is legal too. If failureTask() returns null as an example, it should be obvious that the respective branch simply has no failure task. So if a child is null i would suppose it getting treated as an empty leaf instead.
Is there something I'm missing here? createRootBranch --> custom extension of RootBranch ... bot runs into empty leaf -> return your RootBranch to restart. is there an issue with doing it that way (minus not having the delay)
By linking back to your root, you'll create a cycle, which you don't really want to do. I assume it was done the way it is to try to encourage authors to balance the tree rather than having one which is heavily weighted to a side.
Code (Text): LeafTask emptyLeaf = new LeafTask() {@Override public void execute() {}); Easiest way if you haven't already, instead of making new class.
Yeah thank you i know how to make an empty LeafTask, but my point is using a leaftask that does nothing is pretty much equivalent to a null child, which should get represented as null. In a tree you wouldnt want any nodes with null as content, but rather let the node itself be null.
Tbh I'm torn each way, although I am leaning towards allowing null sub-nodes... @Party @Cloud opinions?
I'm the opposite, I wouldn't mind the idea but I'm definitely leaning in favour of restriction. In my opinion if you're going to have a branch with a null sub-node, that branch should be replaced by a leaf.
