fix: embed woff2 & inlined (data: URI) web fonts#2
Merged
Conversation
Self-contained decks that ship their fonts as
`url("data:font/woff2;base64,…")` — resolved at runtime from a UUID
placeholder in the raw CSS — came out with the font *tagged* but not
embedded: font resolution parsed only the raw <style>/<link> text (still
the unresolved placeholder) and skipped woff2 outright. PowerPoint then
substituted a fallback and bold text rendered as a mismatched faux-bold
("all the fonts are broken").
Font resolution now also reads the LIVE CSSOM for faces inlined as
data:/blob:, decodes woff2 to TrueType (via fonteditor-core's woff2
module — no new dependency) so PowerPoint/dom-to-pptx can read it, and
embeds the real Regular and Bold faces. The existing Node-side parse of
CDN <link>/<style> stylesheets is kept as-is (a plain-UA fetch returns
the FULL font, not the browser's per-script unicode-range subsets), and
CSSOM subset/unicode-range faces are skipped for the same reason. Only
families actually used by the slide text are embedded.
Verified end-to-end with PowerPoint COM render: a woff2 data:-URI deck
(Pretendard) now embeds full Regular+Bold with Korean coverage and
renders faithfully; the CDN-font sample (Noto Sans KR) still embeds the
full 5.9 MB face.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
mmoollee101-lab
added a commit
that referenced
this pull request
Jul 16, 2026
Retrospective PDCA completion report for the fix landed in PR #2 (main 9b022ff): embedding woff2 and inlined data:-URI web fonts. Records root cause, the two-source resolver design, acceptance criteria (6/6), PowerPoint-COM visual verification, and the CDN-subset regression caught during development. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
mmoollee101-lab
added a commit
that referenced
this pull request
Jul 16, 2026
Guards the two failure modes fixed in #2 so they can't silently return: R1 a woff2 font inlined as a data: URI must be embedded (offline) R2 a CDN font must embed FULL, not a unicode-range subset (network-gated) node:test suite drives the public API (convertHtmlToPptx) end-to-end and asserts on the output .pptx bytes: ppt/fonts non-empty, embeddedFontLst has Regular AND Bold, every embedded EOT's cmap covers Korean, and a byte-size floor rejects the 31 KB subset trap. Zero new dependencies (node:test + jszip + fonteditor-core, all already present). The R1 fixture is a self-contained 5.6 KB HTML deck inlining a ~1.7 KB Pretendard subset (OFL-1.1, attributed in test/fixtures/NOTICE.md), produced by scripts/gen-font-fixture.js. R2 reuses samples/sample.html (Noto Sans KR) and skips loudly when the CDN is unreachable. Runs via `npm run test:fonts`, wired into CI as a dedicated `font-embed` job that launches real Chromium (the fast lint/load matrix stays browser- free). Verified: both cases pass, and neutralizing the data:-URI embed makes R1 fail with "no embedded fonts (ppt/fonts empty)". Includes the full PDCA record (plan/design/analysis/report) for the feature. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Problem
Re-converting a self-contained HTML deck (fonts shipped as
url("data:font/woff2;base64,…"), resolved at runtime from a UUID placeholder) produced a.pptxwhere every font was broken: each run was taggedtypeface="Pretendard"butppt/fontswas empty, so PowerPoint substituted a fallback and bold text rendered as a mismatched faux-bold.Two root causes:
<style>/<link>text — where the src is still an unresolved placeholder (a bare UUID) — so it found nothing to embed.Fix
data:/blob:(what self-contained decks actually end up with after their loader runs).fonteditor-core's built-in woff2 module (no new dependency) so PowerPoint / dom-to-pptx — both woff2-blind — can embed it.<link>/<style>unchanged: a plain-UA fetch returns the full font, not the browser's per-scriptunicode-rangesubsets (a subset would embed as a tiny face missing most glyphs). CSSOM subset faces are skipped for the same reason.Verification (PowerPoint COM render, Pretendard not installed system-wide)
Docs (README, CLI help) and CHANGELOG updated.