Right now, when taking the tree from a running parse, then continuing to run it, and taking the tree again, all nodes are recreated. It might be useful if the parser automatically, when the output buffer reaches a certain length and there's only one active stack, converted all finished nodes to Tree objects and rewrote its stack to refer to those.
Right now, when taking the tree from a running parse, then continuing to run it, and taking the tree again, all nodes are recreated. It might be useful if the parser automatically, when the output buffer reaches a certain length and there's only one active stack, converted all finished nodes to
Tree
objects and rewrote its stack to refer to those.