Jint

Working with Trees ​

Use using Jint.HtmlParser; for native tree types. A Document owns its nodes. Walk ordinary children through FirstChild, NextSibling, ParentNode, or ChildNodes; DocumentElement returns the root element. Native Node objects are CLR objects and require no JavaScript engine.

Creation and mutation ​

Create nodes with the owning document: CreateElement, CreateTextNode, CreateComment, CreateCDataSection, CreateProcessingInstruction, CreateAttribute and CreateDocumentFragment. Use CreateElementNS and CreateAttributeNS for explicit namespaces; Namespaces supplies common URIs. CDATA creation requires an XML document.

AppendChild, InsertBefore, ReplaceChild, RemoveChild and ReplaceChildren mutate the tree. Appending a fragment moves its children and empties the fragment. Element.GetAttribute returns null for a missing attribute; SetAttribute and RemoveAttribute update it. Their *NS variants distinguish namespace URI and local name. An Attr has an owner element and is not an ordinary child node.

CloneNode(deep: true) copies a subtree. ImportNode(source, deep: true) copies it into another document; AdoptNode(source) moves ownership while retaining native identity. Keep all tree access synchronized while performing these operations. Normalize merges adjacent ordinary text nodes and repairs live endpoints.

Mutation subscriptions ​

Observe changes with Document.ObserveMutations. The standalone subscription queues records for explicit draining; it does not schedule a JavaScript callback or a browser microtask.

csharp
var document = MarkupParser.ParseXml("<root/>");
var root = document.DocumentElement!;
using var subscription = document.ObserveMutations(root, new MutationObserverOptions
{
    ChildList = true,
    Attributes = true,
    AttributeOldValue = true,
    Subtree = true
});
var item = document.CreateElement("item");
root.AppendChild(item);
item.SetAttribute("id", "one");
foreach (var record in subscription.TakeRecords())
{
    Console.WriteLine(record.Kind);
}

This prints ChildList and Attributes. Enable the record categories you need, and Subtree to include descendants. AttributeOldValue and CharacterDataOldValue request previous values; AttributeFilter restricts attribute names. TakeRecords() returns an independent, read-only record collection and empties the queue; referenced nodes remain mutable. Disconnect() removes registrations; disposing the subscription also releases it. Use TakeRecordsForDelivery() when implementing a delivery cycle that must also clear transient registrations associated with removed subtrees.

Live ranges ​

DomNodeIdentity represents a node or an attribute. BoundaryPoint combines that identity with an unsigned offset: UTF-16 units for character data, child positions for containers. Document.CreateRange() creates a live DomRange; endpoints repair as the tree changes.

csharp
var document = MarkupParser.ParseXml("<root/>");
var text = document.CreateTextNode("abcd");
document.DocumentElement!.AppendChild(text);
var range = document.CreateRange();
range.SelectNodeContents(new DomNodeIdentity(text));
text.ReplaceData(1, 1, "B");
Console.WriteLine(range.GetText()); // aBcd

This prints aBcd. Set individual endpoints with SetStart/SetEnd, or choose a node or its contents. Collapse, point/boundary comparisons and IntersectsNode support selection logic. CloneContents copies selected content; ExtractContents moves it; DeleteContents removes it. InsertNode and SurroundContents edit at the selected boundaries and apply DOM validity checks. Invalid operations raise DomException, whose Name identifies the DOM error.

DomStaticRange records endpoints without following subsequent edits; use IsValid() to check them. It does not freeze the referenced nodes. DomRange.Detach() is a compatibility no-op.

Iterators and tree walkers ​

DomNodeIterator walks tree order and repairs its reference when nodes are removed. DomTreeWalker provides parent, child, sibling and forward/backward movement with a mutable Current identity. Both take a DOM node-type bitmask and an optional synchronous TraversalFilter per movement call:

csharp
var document = MarkupParser.ParseXml("<root><item/></root>");
var iterator = new DomNodeIterator(new DomNodeIdentity(document), whatToShow: 1);
while (iterator.Next(filter: null) is { } identity)
{
    Console.WriteLine(((Element) identity.Node!).LocalName);
}

A mask of 1 selects elements; uint.MaxValue permits all types. Filters return 1 for accept, 2 for reject or 3 for skip. Recursive filtering raises InvalidStateError. Traversal remains an ordinary tree walk: template contents and shadow trees are separate roots that must be traversed explicitly.

Templates and shadow roots ​

A template element's TemplateContent holds its detached content fragment. An attached shadow root has its own descendant tree. Element.AttachShadow(new ShadowRootInit(...)) returns that native root, including in closed mode; OpenShadowRoot exposes only open roots. Attachment performs no Browser custom-element reactions. See Serialization for explicit inclusion.

Selector matching and browser CSSOM helpers are internal integration APIs. Use native traversal or the public XPath API for standalone tree queries.