I've addressed those comments, and added some additional tests which verify that the mask is tied to the handle and not the directory.
If necessary I can squash the two commits that add tests together, for a cleaner git history.
Please let me know if there are any additional concerns or comments :thumbsup: