Skip to content

2.0.0-beta.3

Pre-release
Pre-release

Choose a tag to compare

@hexenq hexenq released this 17 Oct 09:02
2.0.0-beta.3
f9603ff

Released at 2026-10-17

Bug Fixes

  • Preserve kana tokens when their written form and reading have different lengths, avoiding broken annotations.
  • Keep kana combinations, sokuon, and pronounced long vowels together when romanizing tokens and annotations.
  • Preserve pronunciation information and syllabic n context across conversion modes, including mixed kanji and kana annotations.
  • Use an apostrophe for sokuon that cannot double a following consonant, instead of adding a tsu syllable or repeating a vowel.

Security

  • Update source-map-js in development dependencies to address a denial-of-service vulnerability (#132).

Note: Romanization output and annotation grouping change in this release. Unmatched sokuon use an apostrophe in all three systems, for example あっ → a' and っい → 'i.