Crate marki_parse 

A fast, zero-copy CommonMark parser with SIMD-accelerated scanning.

marki-parse parses Markdown into structured Section and Inline elements, borrowing directly from the input string with no intermediate allocations for text content.

Quick start

use marki_parse::MarkdownFile;

let md: MarkdownFile<'_> = MarkdownFile::parse("# Hello\n\nSome **bold** text.");
for section in &md.sections {
    println!("{section:?}");
}

CRLF input

The parser operates on LF (\n) line endings. For input that may contain \r\n, call normalize first — it returns the input borrowed when no \r is present (zero cost):

let input = "# Hello\r\nWorld";
let normalized = marki_parse::normalize(input);
let md: marki_parse::MarkdownFile<'_> = marki_parse::MarkdownFile::parse(&normalized);

Accessing inline elements

Inline elements are stored in a flat pool for cache efficiency. Use MarkdownFile::inlines and MarkdownFile::item_spans (or index with InlineSpan / SpanSlice) to retrieve them:

use marki_parse::{MarkdownFile, Section};

let md: MarkdownFile<'_> = MarkdownFile::parse("Hello **world**");
if let Some(Section::Paragraph { content }) = md.sections.first() {
    for inline in md.inlines(*content) {
        println!("{inline:?}");
    }
}

Structs

InlineSpan

A range of inline elements stored contiguously in the inline pool.

MarkdownFile

A parsed Markdown document.

SpanSlice

A range of InlineSpan elements stored contiguously in the span pool. Used by list sections to avoid per-list Vec<InlineSpan> heap allocations.

Enums

Inline

An inline element within a Markdown block.

OrderedListDelimiter

The delimiter used after the number in an ordered list item (1. vs 1)).

Section

A block-level element of a Markdown document.

SpecialChar

Functions

normalize

Normalize line endings for parsing. Converts \r\n to \n and bare \r (classic Mac) to \n. Returns the input borrowed if no carriage returns are found.