fix: preserve false stylesheet booleans

[baseballyama]

Summary

Font toggles with val="0" or val="false" now stay disabled through load/save, and explicit false XF flags survive serialization.

Motivation

Closes #209

The font schema treated b, i, strike, outline, shadow, condense and extend as presence-only markers, ignoring their val attributes. SpreadsheetML's CT_BooleanProperty uses a boolean attribute with a default of true; the existing nested-boolean schema already models this distinction. The XF writer separately discarded false flags through truthiness checks.

The regression fixture uses the report's stylesheet, with a differential font added to exercise the shared font schema. XML is preserved semantically, not byte-for-byte: true font values serialize as val="1".

Changes

• Parse all seven font toggles as booleans, preserving absent elements and defaulting bare elements to true.
• Preserve explicit false values in regular/differential fonts and both XF tables.
• Add a patch changeset and correct the schema comments describing font toggles.

Testing

• Regression: pnpm test tests/styles/issue-209.test.ts — 45 failures before the fix, all 75 cases pass afterward.
• Covers all four XML boolean spellings, bare/absent elements, whitespace, invalid values, both XF tables, normal/streaming loading, editing and saving. Input and saved XLSX pass OPC/XSD/semantic validation.
• Full suite: 3,271 tests pass; performance gate: 16 tests pass.
• pnpm lint, pnpm typecheck, pnpm knip, pnpm build, pnpm size, pnpm check:attw and pnpm check:consumer pass.

Breaking changes

None

Checklist

• I have read CLAUDE.md and followed the project's conventions.
• I have added or updated tests for the change.
• I have added or updated documentation where user-visible behavior changed.
• If this is a user-visible change, I have run pnpm changeset and committed
  the result.
• If this is a breaking change, I have flagged it above and the changeset is
  marked accordingly.
• I have re-read my own diff and removed dead code, debug prints, and stale
  comments.
• If I used an LLM to draft this PR, I have verified each change myself, this
  PR represents real work that warrants a maintainer's review, and I am
  willing to defend each line in review.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.