Read hdmx device records at their padded size

[Likheet]

Each hdmx device record is padded to sizeDeviceRecord bytes (a multiple of four), but the struct only read 2 + numGlyphs bytes per record. When numGlyphs + 2 is not a multiple of four, every record after the first was read from the wrong offset.

With Arial (4651 glyphs, 3 bytes of padding), font.hdmx.records.map(r => r.pixelSize) gave 11, 0, 6, 9, 6, 15, ... instead of 11, 12, 13, 15, 16, 17, ....

This skips the padding after each record. Spec: https://learn.microsoft.com/typography/opentype/spec/hdmx

None of the fonts in test/data has an hdmx table, so the test builds a small font in memory. npm test passes and the changed files are Prettier-clean.

[Copilot]

🟢 Approval recommended

The focused parser fix matches the record format and includes adequate regression coverage.

0 open findings

What changed in this PR

Fixes hdmx parsing so padded device records remain correctly aligned.

Changes:

• Skips per-record padding using sizeDeviceRecord.
• Adds an in-memory regression test.
• Documents the fix in the changelog.

File	Description
src/​tables/​hdmx.js	Consumes record padding.
test/​device_metrics.js	Tests padded records.
CHANGELOG.md	Records the fix.

🧠 Review effort: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.