RXerces

A Ruby XML library with a Nokogiri-compatible API, powered by Apache Xerces-C instead of libxml2.

Overview

RXerces provides a familiar Nokogiri-like interface for XML parsing and manipulation, but uses the robust Apache Xerces-C XML parser under the hood. This allows Ruby developers to leverage Xerces-C's performance and standards compliance while maintaining compatibility with existing Nokogiri-based code.

Features

✅ Nokogiri-compatible API
✅ Powered by Apache Xerces-C
✅ Parse XML documents
✅ Navigate and manipulate DOM trees
✅ Read and write node attributes
✅ Query nodes with XPath (basic support)
✅ Serialize documents back to XML strings

Installation

Prerequisites

You need to have Xerces-C installed on your system:

macOS (Homebrew):

brew install xerces-c

Ubuntu/Debian:

sudo apt-get install libxerces-c-dev

Fedora/RHEL:

sudo yum install xerces-c-devel

Install the Gem

Add this line to your application's Gemfile:

gem 'rxerces'

And then execute:

bundle install

Or install it yourself as:

gem install rxerces

Usage

Basic Parsing

require 'rxerces'

# Parse XML string
xml = '<root><person name="Alice">Hello</person></root>'
doc = RXerces.XML(xml)

# Access root element
root = doc.root
puts root.name  # => "root"

Nokogiri Compatibility

RXerces provides a Nokogiri module for drop-in compatibility:

require 'rxerces'

# Use Nokogiri syntax
doc = Nokogiri.XML('<root><child>text</child></root>')
puts doc.root.name  # => "root"

# Classes are aliased
Nokogiri::XML::Document == RXerces::XML::Document  # => true

Working with Nodes

# Parse XML
xml = <<-XML
  <library>
    <book id="1" title="1984">
      <author>George Orwell</author>
      <year>1949</year>
    </book>
    <book id="2" title="Brave New World">
      <author>Aldous Huxley</author>
      <year>1932</year>
    </book>
  </library>
XML

doc = RXerces.XML(xml)
root = doc.root

# Get attributes
book = root.children.find { |n| n.is_a?(RXerces::XML::Element) }
puts book['id']     # => "1"
puts book['title']  # => "1984"

# Set attributes
book['isbn'] = '978-0451524935'
puts book['isbn']   # => "978-0451524935"

# Get text content
author = book.children.find { |n| n.name == 'author' }
puts author.text    # => "George Orwell"

# Set text content
author.text = "Eric Arthur Blair"
puts author.text    # => "Eric Arthur Blair"

Navigating the DOM

# Get all children
root.children.each do |child|
  puts "#{child.name}: #{child.class}"
end

# Find specific elements
books = root.children.select { |n| n.is_a?(RXerces::XML::Element) && n.name == 'book' }
books.each do |book|
  puts "Book ID: #{book['id']}"
end

Serialization

# Convert document back to XML string
xml_string = doc.to_xml
puts xml_string

# or use to_s
puts doc.to_s

XPath Queries

RXerces supports XPath queries using Xerces-C's XPath implementation:

xml = <<-XML
  <library>
    <book>
      <title>1984</title>
      <author>George Orwell</author>
    </book>
    <book>
      <title>Brave New World</title>
      <author>Aldous Huxley</author>
    </book>
  </library>
XML

doc = RXerces.XML(xml)

# Find all book elements
books = doc.xpath('//book')
puts books.length  # => 2

# Find all titles
titles = doc.xpath('//title')
titles.each do |title|
  puts title.text.strip
end

# Use path expressions
authors = doc.xpath('/library/book/author')
puts authors.length  # => 2

# Query from a specific node
first_book = books[0]
title = first_book.xpath('.//title').first
puts title.text  # => "1984"

Note on XPath Support: Xerces-C implements the XML Schema XPath subset, not full XPath 1.0. Supported features include:

Basic path expressions (/, //, ., ..)
Element selection by name
Descendant and child axes

Not supported:

Attribute predicates ([@attribute="value"])
XPath functions (last(), position(), text())
Comparison operators in predicates

For more complex queries, you can combine basic XPath with Ruby's select and find methods.

API Reference

RXerces Module

RXerces.XML(string) - Parse XML string and return Document
RXerces.parse(string) - Alias for XML

RXerces::XML::Document

.parse(string) - Parse XML string (class method)
#root - Get root element
#to_s / #to_xml - Serialize to XML string
#xpath(path) - Query with XPath (returns NodeSet)

RXerces::XML::Node

#name - Get node name
#text / #content - Get text content
#text= / #content= - Set text content
#[attribute] - Get attribute value
#[attribute]= - Set attribute value
#children - Get array of child nodes
#xpath(path) - Query descendants with XPath

RXerces::XML::Element

Inherits all methods from Node. Represents element nodes.

RXerces::XML::Text

Inherits all methods from Node. Represents text nodes.

RXerces::XML::NodeSet

#length / #size - Get number of nodes
#[] - Access node by index
#each - Iterate over nodes (Enumerable)
#to_a - Convert to array

Development

Building the Extension

bundle install
bundle exec rake compile

Running Tests

bundle exec rspec

Running Tests with Compilation

bundle exec rake

Implementation Notes

Uses Apache Xerces-C 3.x for XML parsing
C++ extension compiled with Ruby's native extension API
XPath support is basic (full XPath requires additional implementation)
Memory management handled by Ruby's GC and Xerces-C's DOM

Differences from Nokogiri

While RXerces aims for API compatibility with Nokogiri, there are some differences:

Parser Backend: Uses Xerces-C instead of libxml2
XPath: Basic XPath support (returns empty NodeSet currently)
Features: Subset of Nokogiri's full feature set
Performance: Different performance characteristics due to Xerces-C

Contributing

Fork it
Create your feature branch (git checkout -b my-new-feature)
Commit your changes (git commit -am 'Add some feature')
Push to the branch (git push origin my-new-feature)
Create new Pull Request

License

MIT License - see LICENSE file for details

Credits

Built with Apache Xerces-C
API inspired by Nokogiri

Misc

This library was almost entirely written using AI (Claude Sonnet 4.5). It was mainly a reaction to the lack of maintainers for libxml2, and the generally sorry state of that library in general. Since nokogiri uses it under the hood, I thought it best to create an alternative.

Copyright

Author

Daniel J. Berger