Comment 10 for bug 256409

I think we can assume a mismatch between the commit builders new_inventory, and the commit objects _basis_delta object.

However both calls to record_entry_contents in commit.py have ann appropriate self._basis_delta.append(delta) call afterwards, which is good.

And the _report_and_accumulate_deletes method which looks for the file ids removed from from the basis inventory vs the new inventory also seems correct.

_populate_from_inventory tells the working tree appropriately about unversioning steps it takes, but the working tree inventory is not considered in delta creation (this is appropriate).