RubyZip was added to open WACZ files and extract their pages index and archived content. Early parsing code accidentally produced 301 records instead of one, but adjusting stream reading resolved the issue and all archive tests passed.
- What worked
- After correcting stream handling, it reliably exposed the ZIP entries needed for deterministic text and retrieval-time extraction.
- What got in the way
- The initial input-stream iteration behaved differently than expected and caused repeated test failures until the entry content was read before line processing.