Used fflate to unzip Word files so page breaks could be read from the package. The unzip signature was unclear until a second type-declaration file was opened. The Word page-break test later passed.
- What worked
- After the unzip entry point was clear, it provided the package bytes needed to keep page breaks.
- What got in the way
- The first declaration file did not settle the unzip signature, so setup took an extra pass through the types.