The canonical, dependency-free DOM and HTML parsing component for Broiler, targeting .NET 10.
It contains:
Broiler.Dom: document, node, element, attribute, mutation, range, and traversal primitives.Broiler.Dom.Html: HTML tokenization, document and fragment parsing, tree building, and serialization.
Every push and pull request builds both libraries on Linux and Windows, runs both test suites, and packs the two packages to prove they still build. Both legs run the same platform-neutral code deliberately: a tokenizer is where two hosts disagree about newlines, string comparison and culture, and the character-reference decoding this component hands to the platform is a different platform on each.
Publish pushes to NuGet.org only, with automatic preview versions; CI packs every package and verifies a consumer restore on every run without pushing. See CI, packages, and releases for configuration, credentials, and local packaging commands.
Both bodies that have stewarded the DOM and HTML — the WHATWG and the W3C — operate royalty-free patent policies. Both bind participants over Essential Claims and both let a participant exclude specific patents during a disclosure window, so neither is a statement that no relevant claim exists.
That, and what this repository does and does not contain, is recorded in the DOM IP and licensing register, where every row is now decided. Decided is not cleared: no lawyer reviewed it, patent-freedom is not claimed, and no freedom-to-operate determination has been made. Both patent policies let a participant exclude patents during a disclosure window, which is what keeps a royalty-free commitment from being a finding that no claim exists.
One finding is worth repeating outside the register. The HTML Standard's named
character reference table — roughly 2,231 entries, and the largest thing an HTML
parser might copy — is not in this repository. Two call sites hand the work to
System.Net.WebUtility, and a guard test fails the build if a transcribed table
ever appears. Neither library takes a package reference, no third-party
conformance suite is committed, and no specification text is reproduced.
This component is a tokenizer, a tree builder and a serializer. It is not a browser engine and nothing here may describe it as one, or as HTML5-conformant.
This is first-preview software. Its API and behavior may change without compatibility guarantees. Substantial implementation work was AI-assisted. Human-review approval is revision-scoped; consult HUMAN_REVIEW.md for the reviewed revision and conditions before describing the current checkout as approved.
Broiler.DOM is an independent Broiler component. It interoperates with Broiler.HTML, whose rendering lineage comes from HTML Renderer, but it must not be represented as an official HTML Renderer component or as endorsed by that project's contributors.
Broiler.Dom is the only mutable tree and owns engine-neutral DOM algorithms;
Broiler.Dom.Html owns parsing and serialization. Renderer geometry and JavaScript
wrappers remain in HtmlBridge because they depend on computed style, layout, or the
script runtime.
The current HtmlBridge pipeline consumes this canonical model directly. Bridge-owned JavaScript wrappers do not constitute a second DOM tree, and bridge compatibility or public-surface work is owned by the main Broiler repository rather than this component.
dotnet build Broiler.Dom.slnx
dotnet test Broiler.Dom.slnx- Current roadmap — the remaining component release gate
- Human-review record — revision-scoped preview decision
Broiler.DOM is licensed under the Apache License 2.0. Third-party material, if present, retains the license identified with that material. The license provides the software on an “AS IS” basis, without warranties or conditions.