GFM Markdown parsing and formatting added (
a7f0c48)
Biome now supports GitHub Flavored Markdown in its markdown parser, formatter, and generated syntax/node factories. That unlocks structured handling for GFM features like strikethrough, tables, and task list items, making Markdown support much broader and more standards-aligned.
Type inference avoids eager generic expansion (
20e513a)
The type-inference path no longer fully infers imported generic declarations just to apply type arguments, which should materially improve lint performance on large type-heavy libraries. The commit also adds regression coverage for several affected rules and issue cases.
Token lookup by offset is faster (
ea20e5a)
Rowan’s offset lookup now uses a more direct green-tree-based path instead of scanning child elements, with new tests covering boundary and zero-width cases. This improves linting performance, especially for large CSS and JSON files.
Grit named import matching fixed (23aad6d)
Grit plugins can now capture and inspect multiple named import specifiers correctly. This fixes a regression where queries over imports could miss or mishandle matched specifier lists.
Other misc changes
- Dependency bump:
unicode-widthupdated to 0.2.2. - Markdown parser configuration gained a
gfmtoggle defaulting to enabled. - Console replacement-character handling now treats control characters as invalid graphemes.
- Formatter-development guidance and review checklist were expanded.
- Added/updated regression tests and snapshots for the above fixes.