Hm, that's strange – the resulting HTMLs from huffdic compressed e-books from HarperCollins are twice as big as they should be. I looked at the generated HTMLs and found out that after they should end, the books start again with the table of contents like:
Title page(s)
Table of Contents
Book content
Book ending
Table of Contents (this and the following content shouldn't be there)
Book content
Book ending
You can test this with the free e-book
Flight of the Nighthawks by Raymond E. Feist.