The DOM

jsdom is the layer between the HTML tree and the JavaScript engine.

Nodes and trees

The full node hierarchy is implemented: documents, elements, text, comments, document fragments, doctypes and processing instructions, with the standard mutation methods, Range-free tree operations, cloneNode, importNode, adoptNode and normalize. Cloning wires inline event handlers on the copy.

Events

Dispatch follows the standard: capture, target and bubble phases, stopPropagation and stopImmediatePropagation, once, passive (defaulting to passive for the listener types the spec names), AbortSignal removal and legacy webkit aliases. Activation behaviour is implemented for the elements that have it (form submit and reset, links, summary, label), acting on the nearest activatable ancestor only.

Live collections

getElementsByTagName, childNodes, document.images, forms, links and friends return live collections. DOMTokenList supports indexed access through the engine's indexer hook, and length is a prototype getter so own-key enumeration is correct.

Reflection

HTML attribute reflection (el.href, el.tabIndex, el.hidden, ...) is generated from WebIDL into webreflect. One accessor pair, parameterised by row, serves every reflected property, so adding an element is a table change.

Frames

An <iframe> loads a real document. Frame documents get their own cascade, their own realm and their own base URL, and fire their own load event, which the parent's load waits for. XML frames parse as XML.

Parsing and serialisation

DOMParser, XMLSerializer, innerHTML, outerHTML and insertAdjacentHTML share the HTML and XML parsers. <base href> sets the document base URL, and the document's characterSet reflects the decoded encoding.

Not implemented on purpose

Interfaces for features that do not exist here (canvas, workers, event sources) are not exposed. A page that feature-detects them gets the truth. See Limits and roadmap.