This content is for v2.x version. Switch to the Stable version.

Roadmap

The v2.x line builds on the stable, round-trip-safe DOM from v1.x. Its headline goal is native XPath querying, plus supporting ergonomics for selecting and extracting data from a parsed tree.

Headline: XPath support Planned

Today you navigate the tree with structural helpers such as find_first_child, get_elements_by_tag_name and find_first_by_attribute. v2 aims to add a single expression-based entry point so you can locate nodes with familiar XPath syntax instead of chaining calls.

The target API is a query_xpath method on XmlDocument:

// Target API — not yet available.
let titles = document.query_xpath("//book[@id='bk101']/title")?;
for node_id in titles {
    let title = document.get_element(node_id)?;
    println!("{}", title.get_text_content());
}

Planned expression coverage

Capability	Example	Status
Child & descendant steps	`/catalog/book`, `//title`	Planned
Attribute predicates	`book[@id=‘bk101’]`	Planned
Positional predicates	`book[1]`, `book[last()]`	Planned
Wildcards	`/catalog/`, `@`	Planned
Namespace-qualified steps	`//t:book` via a bound prefix map	Exploring
Text & node tests	`//title/text()`	Exploring

Design goals

Zero-copy over the existing DOM — XPath evaluation returns NodeId values into the current tree, with no separate document model.
Namespace-aware — expressions resolve prefixes through the same NamespaceDeclaration scope used elsewhere, keeping results round-trip-safe.
No unsafe — consistent with the rest of the crate.
Incremental — child/descendant axes and attribute predicates ship first; richer axes and functions follow.

Beyond XPath

Area	Intent	Status
Streaming / SAX-style reader	Event callbacks for very large documents without building a full DOM.	Exploring
Pretty-printing option	Optional indentation on serialize alongside the existing compact output.	Exploring
Broader encoding support	Read non-UTF-8 declarations by transcoding to UTF-8 on parse.	Exploring