Token-Oriented Object Notation Files

   🚀 Features

• benchmarks:
• Add keyed and nested-group datasets covering v4 tabular forms  -  by @johannschopplich (ebe80)
• Corrupt encoded text for structural-validation datasets  -  by @johannschopplich (1de7b)
• Add wilson 95% intervals to accuracy tables  -  by @johannschopplich (ac638)
• encode:
• rawString marker and escapeString export for verbatim replacer output  -  by @yilmazhasan in https://github.com/toon-format/toon/issues/321 (b36f8)

   🐞 Bug Fixes

• Accept the indentSize option named by the spec  -  by @johannschopplich (b41d5)
• benchmarks:
• Normalize non-ISO dates in local time to avoid timezone-dependent day shifts  -  by @johannschopplich (c7b6f)
• Pin faker reference date so regenerated datasets stay reproducible  -  by @johannschopplich (d7d78)
• Split accuracy report into csv-excluded and flat-only tracks  -  by @johannschopplich (911ba)
• Floor grok-4.5 reasoning at low  -  by @johannschopplich (7c792)
• cli:
• Error on ill-formed UTF-8 input in strict mode  -  by @johannschopplich (5f04e)
• decode:
• Strip a byte-order mark and trailing spaces, and stop truncating at a declared [N]  -  by @johannschopplich (d8ca3)
• Keep a quoted colon inside an unquoted key  -  by @johannschopplich (41c92)
• Omit leaf fields that have no cell instead of yielding undefined  -  by @johannschopplich (c2c25)
• Reject a bare token line outside root primitive position  -  by @johannschopplich (ea050)
• encode:
• Reject unpaired surrogates and invalid delimiters  -  by @johannschopplich (db67c)

    View changes on GitHub