Crate asciidoc_parser

Expand description

§AsciiDoc parser for Rust

This is a semantic parser for the AsciiDoc language written in the Rust language.

As of version 0.25.0 (July 2026) this crate is feature-complete, heavily tested, and ready to be used.

Now that the core is in place, I’ll be building the downstream projects described below. Expect the API to evolve slightly in the coming months as those projects mature. I do expect to publish a mature (1.0) release within the year.

§Why do this?

Most of all this is a fun project that exercises different architectural and project design skills from my day job. As part of that work, I write technical standards for the Creator Assertions Working Group in Asciidoc and Antora.

There are a few projects that I’m now starting to build that depend on the parser:

A version of Antora that highlights differences between versions of a spec/document, as in version to version or proposed updates in a pull request.
A version of Antora or similar that shows what portions of a spec are tested/completed/known good. (See the following section on “spec-driven development.”)
A version of Zola, the static site generator that I use for most of my web sites, that accepts Asciidoc formatted text as input. (See Project proposal: Asciidoc support in Zola.)

§Spec-driven coverage

I value high code coverage. Code coverage for this crate is now extremely high (99.5%). But it’s not just code that I’m covering.

In this project, I also employ a technique I call “spec-driven development.” Since I started, that phrase has taken on a different and now more widely used meaning — writing a structured specification up front so that an AI coding agent can implement it, as popularized by tooling such as GitHub’s Spec Kit. That is not what I mean here. In my sense the specification already exists — it’s the AsciiDoc language description — and I’m driving the implementation toward it: not only am I monitoring coverage of the code but also coverage of the spec.

I’ve read the language definition and Asciidoctor’s extensive test suite page-by-page, line-by-line, and written tests to verify that this implementation matches the specification(*) and Asciidoctor’s behavior. This slowed progress considerably, but I believe it has resulted in an implementation that is far more solid than it would otherwise be.

(*) Yes, I’m aware that the Asciidoc language authors consider this a “language description,” not a specification. I’m splitting the difference here.

§Extensions

Extensions – custom block, block macro, inline macro, and similar processors, as provided by the Ruby implementation of Asciidoctor – are not planned for the 1.0 release of this crate. They may be added in a subsequent version.

§No planned support for some AsciiDoc features

The following features are supported in the Ruby implementation of Asciidoctor, on which this project is based, but are not supported – and will likely never be supported – in this crate:

Parsing UTF-16 content is not supported. (UTF-16 documents must be re-encoded to UTF-8 prior to parsing with this crate.)
Document types other than article are not supported. Specifically, features which are enabled for the book doctype are not supported. These include book parts – level-0 section headings in the document body, which are reported via WarningType::Level0SectionHeadingNotSupported (see #800) – and the partintro block style that goes with them (see #794). A few behaviors that Asciidoctor gates on the book doctype are nonetheless implemented incidentally; they should not be relied upon.
The document attribute compat-mode is not supported.
The legacy two-line (or setext) heading syntax — a line of title text underlined by a row of = or - characters — is not supported. It is not part of the AsciiDoc language description, which defines only the single-line (ATX) form (= Document Title, == Section Title); only that form is recognized.
The parser has built-in support for HTML5 rendering similar to what is provided in Asciidoctor. Other back ends could be supported by other crates by implementing the InlineSubstitutionRenderer trait. They will not be directly supported in this crate.
Setting document attributes via the inline attribute entry syntax ({set:name:value} / {set:name!}) is not supported. (Note that this syntax is discouraged and may eventually be removed from the AsciiDoc language documentation.) As a consequence, per-cell table background colors set via the {set:cellbgcolor:...} document attribute are also not supported.
Retrieving include file content via URL is not directly supported. An implementation could implement the IncludeFileHandler trait to provide that behavior.
The shorthand menu syntax ("File > Save") is not supported. Per the AsciiDoc language documentation, it is not on a standards track, so only the menu: macro form is implemented. (The kbd:, btn:, and menu: UI macros themselves are supported.)

§Licenses

The asciidoc-parser crate is distributed under the terms of both the MIT license and the Apache License (Version 2.0).

See LICENSE-APACHE and LICENSE-MIT.

Note that some components and dependent crates may be licensed under different terms; please check the license terms for each crate and component for details.

IMPORTANT: This project is a personal project; it is known to my team at Adobe, but it is not an officially-sponsored project in any way.

§License for AsciiDoc language materials

IMPORTANT: This repository contains a snapshot of the AsciiDoc language description which comes with its own license terms. It is not the purpose of this repository to supplant or replace that description; these documents are here as part of tooling to ensure that this crate follows the language description as closely as possible. Please consult AsciiDoc Language @ Eclipse GitLab for the official language description.

The snapshot lives in ref/asciidoc-lang; see ref/asciidoc-lang/README.md for the exact upstream commit it was taken from and how to refresh it.

The following applies to content in the ref/asciidoc-lang/docs folder:

The user documentation for the AsciiDoc Language, located in the docs/ folder, is made available under the terms of a Creative Commons Attribution 4.0 International License (CC-BY-4.0).

The AsciiDoc Language project as a whole is made available under the terms of the Eclipse Public License v 2.0 (EPL-2.0). See the project LICENSE for the full license text.

Re-exports§

pub use document::Document;
pub use parser::Parser;
pub use parser::ReferenceTime;
pub use parser::SafeMode;

Modules§

attributes: Element attributes are a powerful means of controlling the built-in settings of individual block and inline elements in the AsciiDoc syntax. They can also be used to add supplemental information, such as citation metadata and fallback content, to certain elements.
blocks: Block elements form the main structure of an AsciiDoc document, starting with the document itself.
content: Describes the content of a non-compound block after any relevant substitutions have been performed.
document: Describes the top-level document structure.
parser: The Parser struct and its related structs allow a caller to configure how AsciiDoc parsing occurs and then to initiate the parsing process.
strings: String types that facilitate parsing.
warnings: Describes conditions where a parse result might be unexpected.

Structs§

Span: Represents a subset of the overall UTF-8 input stream.

Traits§

HasSpan: Any syntactic element can describe its location within the source material using this trait.