[ot][spam][crazy] log 2: append-only random writes

[Undiscussed Horrific Abuse, One Victim of Many]

member variables draft. i think this is all the information that was
being replaced by leaf and index iteration.
- leaf count can be updated by summing children when graph is changed.
- height is simply maximum child height + 1, can also update when
graph is changed
- parents is for adding dropped nodes to when branches are made more
shallow. it can be prefix-compared against to find a shared ancestor
when unifying neighboring entries. this shouldn't need an update for
graph changes since the nontrimmed data is immutable.

class Data:
    def __init__(self, offset, data, end = None, height = 0, parents =
[], leaves = 1):
        self.start = offset
        self.data = data
        if end is None:
            self.end = self.start + len(self.data)
        self.height = 0
        self.parents = parents
        self.leaves = 1