The Chinese print edition embeds OFL-licensed subsets of Noto Serif SC. The subset character set is derived from the current Chinese EPUB, while IPA characters not present in Noto Serif SC remain in the separate Noto Sans fallback subset.

The pinned upstream source is ofl/notoserifsc/NotoSerifSC[wght].ttf from Google Fonts commit 2e61f4355afd22b801791b0df176065082423b87, SHA-256 050080d9255a86808f2945bffac582b31ef32bc36411ce29563b4961670c66f9.

To regenerate the 400 and 700 subsets:

python3 -m pip install -r requirements-fonts.txt
python3 scripts/subset-pdf-fonts.py --source-font /path/to/NotoSerifSC[wght].ttf
npm run book:pdf:build

Verify the downloaded source SHA before regeneration. The PDF builder performs a second coverage check, including required spacing glyphs, and fails before publication if any required manuscript character is missing.