Gukhanmun 0.3.1

Released on November 5, 2026.

gukhanmun

• Markdown conversion escapes source text through the re-exported adapter, so the Rust API no longer drops the escapes and character references in the input. [#35, #36]
• HTML and Markdown conversions now escape dictionary readings end to end through the re-exported adapters. [GHSA-j87h-8h88-wm6g]

gukhanmun-cli

• --format text/markdown now preserves escapes and character references in source text both before and after the Hongdown formatting pass. [#35, #36]
• --format text/html and --format text/markdown now escape dictionary readings. [GHSA-j87h-8h88-wm6g]

gukhanmun-core

• Marked annotation-derived text with the new RenderedToken::Annotated variant so that format writers serialize a dictionary reading as literal text instead of treating it as already-valid markup. The variant is a breaking change for code that matches RenderedToken exhaustively; a _ => arm still compiles but silently drops the text. Upgrade gukhanmun-core, gukhanmun-html, and gukhanmun-markdown together. [GHSA-j87h-8h88-wm6g]

gukhanmun-html

• Escaped dictionary readings and other annotation text when writing HTML, including quotes and a guarded first character, so a reading can no longer become an element or an attribute. [GHSA-j87h-8h88-wm6g]
• Stopped treating <script/>, <style/>, and <textarea/> as self-closing so their contents stay raw text, matching browsers. Text after a self-closed raw-text tag inside SVG or MathML is now left unconverted rather than risking conversion inside a script or style body. [GHSA-j87h-8h88-wm6g]

gukhanmun-markdown

• Escaped source text when writing Markdown, so backslash escapes and character references consumed by the parser are restored instead of being re-emitted raw. Interior <, &, [, ], `, *, _, ~, |, and \ can no longer reform HTML, links, images, emphasis, code spans, or tables, and line-start characters can no longer start a heading, list, block quote, or definition list after conversion. Code blocks, autolinks, and HTML blocks remain untouched. Literal & and < in prose now appear as & and < in Markdown output; the rendered text is unchanged. [#35, #36]
• Encoded dictionary readings as literal Markdown text, with U+200B separators that stop GFM from forming an autolink across a reading, so annotation text can no longer form links, emphasis, code spans, tables, or raw HTML. [GHSA-j87h-8h88-wm6g]

gukhanmun-napi

• Markdown output escapes source text. [#35, #36]
• HTML and Markdown output escapes dictionary readings. [GHSA-j87h-8h88-wm6g]

gukhanmun-wasm

• Markdown output escapes source text. [#35, #36]
• HTML and Markdown output escapes dictionary readings. [GHSA-j87h-8h88-wm6g]

@gukhanmun/napi

• Markdown conversion escapes source text. [#35, #36]
• convert(..., "html") and stream("html") escape dictionary readings. [GHSA-j87h-8h88-wm6g]

@gukhanmun/wasm

• Markdown conversion escapes source text. [#35, #36]
• convert(..., "html") and stream("html") escape dictionary readings. [GHSA-j87h-8h88-wm6g]