Releases

v0.17.0

Add opt-in language prefixes for inline code spans with Markly::INLINE_CODE_INFO, expose code metadata through Node#code_info, and provide Node#code_language as a convenient language accessor.
Expose fenced code-block metadata through Node#fence and Node::Fence.
Add Node#next_header, retain Node#next_heading as an alias, and correctly recognize header nodes.
Raise NotImplementedError from the base renderer's unimplemented code-block callback instead of recursing indefinitely.
Remove unused legacy separator helpers from the Ruby renderer.

Update cmark-gfm from upstream, including two table-rendering protections: avoid repeatedly scanning preceding cells, and limit the number of automatically completed cells to prevent denial of service.
Correct end_line source positions for single-line and multi-line HTML blocks.
Fix trailing newlines when rendering inline nodes.
Add support for front matter (CMARK_OPT_FRONT_MATTER): a --- delimited block at the start of a document is captured as a CMARK_NODE_FRONT_MATTER node. The raw content is available via node.string_content and an optional format hint (e.g. "yaml", "toml") via node.fence_info.
Allow : in HTML tag names to support XML namespace prefixes (e.g. <svg:circle>, <xhtml:div>).

Introduced Markly::Renderer::Headings class for extracting headings from markdown documents with automatic duplicate ID resolution. When rendering HTML with ids: true, duplicate heading text now automatically gets unique IDs (deployment, deployment-2, deployment-3). The Headings class can also be used to extract headings for building navigation or table of contents.

Expose Markly::Renderer::HTML.anchor_for method to generate URL-safe anchors from headers.