A common one I see fairly often is opening a big file, creating a “new slice” on a subset of the file and then using the “new slice” and expecting the old large object to be dropped.
Except, the “new slice” is just a reference into the larger slice, so its never marked unused.