> Now that I think more about it, with the current implementation, I'd expect > this behavior: > > * If you delete a node and insert another node instead, the iterator should > go over the child nodes of the new node. I think that's actually reasonable. > * If you insert a node after the one just yielded, the iterator should > still yield the child nodes of the last yielded node and then yield the newly > inserted node and its children. Also not too bad. >
> > So yes, I'll experiment with insertion and add tests if the insertions work > as expected. :-) This turned out to be more complicated than I thought. I should have looked at the code before writing this. ;-) I'm going to open a new thread on the further development of the iterator.
