These methods take the current additions and deletions into account; on success (or basic not-found), they add and remove rrs found in their respective buffers of the diff; i.e. added data is added to the result, removed data is removed. In effect, this returns the data as if the diff would have been committed. These only work on basic data, and do not take special find() processing into account. Therefore, any 'other' result is directly returned (and left to be handled by the caller).

This needed one other change; if you add data which is also in the deletions list, instead of adding it to the additions, it is removed from the deletions (and vice versa). This has the additional benefit of making diffs smaller if they contain removal and addition of the same data.