Right now, for example, if you were to do a find() operation in Biggie, it would have to start from the root, but in some cases it would be better to start operations in subtrees to improve performance and have more flexible functionality.

In addition, this also fixes a few issues when running jstest that makes the storage engine crash, such as retrieving the size of the data, truncating the collection when using --repeat in resmoke.py and the merge() adding children to the incorrect node