fix(ast): preserve positions and fence data through rawHtml reparse

[Princesseuh]

Closes #322.

Changes

• Preserve the original source and root span through the raw-HTML reparse, including UTF-16 offsets and MDX component overrides used by optimizeStatic.
• Share tokenizer-level provenance tracking for positions and fence metadata: serialized byte ranges identify original tokens, and parser handles retain their association through relocation. No HTML markers are added for provenance, and elements are not matched by content or output order.
• Preserve data.lang and data.meta on surviving original code-fence elements, without attaching fence metadata to user-supplied raw HTML or parser-created formatting clones.
• Retain Markdown-derived positions and self-contained raw-block spans where the source association is unambiguous. Split/coalesced nodes, uncertain raw contexts, and nodes parsed within a raw block remain unpositioned. Missing source positions are not invented; arena IDs are rebuilt.

Updated against current main, retaining its SVG/MathML behavior. The previous direct-token feeding approach is replaced with serialization/tokenization so swallowed tags and raw-text contexts keep their existing parsing semantics. Includes separate changesets and updated API/website documentation.

Validation

• cargo test --all
• cargo test -p satteri-ast --no-default-features --features from-html
• Full npm package suite: 3,395 passed, 23 expected failures, one skipped
• Release native binding and TypeScript builds; TypeScript checking
• pnpm lint, all-target Clippy, and formatting checks

Regressions cover Markdown/MDX HAST plugin observability and compilation, repeated/lookalike raw HTML, foster-parented fences, swallowed or malformed tags, formatting clones, nested MDX, source positions, UTF-16 offsets, coalesced text, and conservative raw-block spans.

[codspeed]

Merging this PR will not alter performance

✅ 38 untouched benchmarks
⏩ 10 skipped benchmarks¹

---

_{Comparing fix/rawhtml-root-position (5607aa3) with main (0c2c869)}

Footnotes

1. 10 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩