Module markdown 

The markdown inline scanner.

Implements the restricted subset in docs/design.md “Markdown” — **bold**, *italic* / _italic_, `code`, ~~strike~~, [label](url), and backslash escapes. Block constructs (fenced code, > quotes) are recognised by split_blocks before this scanner runs; everything else CommonMark defines is deliberately absent.

Why hand-written rather than pulldown-cmark. Three reasons: it has no inline-only mode, so we would be filtering a block-level event stream and fighting CommonMark’s block rules to suppress headings, thematic breaks and setext underlines — all of which occur constantly in real chat prose (# 1, ---, ====); its event stream would still need converting into byte-ranged Spans, which is most of the work; and a scanner for six constructs is small enough to be exhaustively tested and predictable on pathological input.

What is deliberately not supported, and why: headings (# opens far too many ordinary chat lines), images (![]() — inline media has a server-validated upload/download pipeline and must not be bypassable by an arbitrary URL), tables, raw HTML, reference links, footnotes, thematic breaks, setext headings, and autolinking (that is crate::linkify’s, which also supplies the scheme list a [label](url) is checked against).

Structs

SourceSpan

One tintable region of source text, for the compose box.

Enums

RawBlock

A block-level piece of a message body.

Functions

parse_inline

Parse inline markdown into styled text.

parse_inline_with

parse_inline, with [label](url) links checked against links rather than the default scheme list.

scan_delims

Locate the markdown delimiters in src that will actually be consumed, reporting ranges in the source.

scheme_allowed

Whether a [label](url) link may point at url under the default scheme list.

split_blocks

Split a message body into block-level pieces.