moxml 0.1.14 → 0.1.16

data/lib/moxml/node.rb

@@ -16,8 +16,14 @@ module Moxml
 
     def initialize(native, context)
       @context = context
-      # @native = adapter.patch_node(native)
       @native = native
+      @parent_node = nil
+    end
+
+    # Update native reference after identity-changing operations
+    # (e.g., LibXML doc.root= creates a new Ruby wrapper)
+    def refresh_native!(new_native)
+      @native = new_native
     end
 
     def document
@@ -29,9 +35,10 @@ module Moxml
     end
 
     def children
-      NodeSet.new(
+      @children ||= NodeSet.new(
         adapter.children(@native).map { adapter.patch_node(_1, @native) },
         context,
+        self,
       )
     end
 
@@ -46,29 +53,40 @@ module Moxml
     def add_child(node)
       node = prepare_node(node)
       adapter.add_child(@native, node.native)
+      # Refresh native in case adapter changed identity (e.g., LibXML doc.root=)
+      refreshed = adapter.actual_native(node.native, @native)
+      node.refresh_native!(refreshed) if refreshed && refreshed != node.native
+      node.parent_node = self
+      invalidate_children_cache!
       self
     end
 
     def add_previous_sibling(node)
       node = prepare_node(node)
       adapter.add_previous_sibling(@native, node.native)
+      invalidate_parent_children_cache!
       self
     end
 
     def add_next_sibling(node)
       node = prepare_node(node)
       adapter.add_next_sibling(@native, node.native)
+      invalidate_parent_children_cache!
       self
     end
 
     def remove
+      invalidate_parent_children_cache!
       adapter.remove(@native)
+      invalidate_children_cache!
       self
     end
 
     def replace(node)
       node = prepare_node(node)
+      invalidate_parent_children_cache!
       adapter.replace(@native, node.native)
+      invalidate_children_cache!
       self
     end
 
@@ -219,6 +237,11 @@ module Moxml
       klass.new(node, context)
     end
 
+    # Internal: Set the parent node for cache invalidation tracking.
+    # Called by NodeSet, Document, Element when establishing parent-child
+    # relationships. Public to allow cross-class usage within Moxml internals.
+    attr_writer :parent_node
+
     protected
 
     def adapter
@@ -229,6 +252,18 @@ module Moxml
       context.config.adapter
     end
 
+    # Invalidate cached children. Called by mutation methods
+    # and by Element attribute/namespace caches.
+    def invalidate_children_cache!
+      @children = nil
+    end
+
+    # Invalidate parent's cached children when this node
+    # is removed/replaced from its parent's child list.
+    def invalidate_parent_children_cache!
+      @parent_node&.invalidate_children_cache!
+    end
+
     private
 
     def prepare_node(node)
@@ -260,54 +295,8 @@ module Moxml
       return options[:declaration] if options.key?(:declaration)
       return options.fetch(:declaration, false) unless is_a?(Document)
 
-      # For Document nodes, check both wrapper flag and native state
-      # Wrapper flag is set by Context.parse for parsed documents
-      # Native state reflects programmatic changes (e.g., add/remove)
-
-      adapter_name = adapter.to_s.split("::").last
-
-      case adapter_name
-      when "Nokogiri"
-        # Nokogiri: if @xml_decl is explicitly set, use that state
-        # Otherwise, trust wrapper flag (for parsed documents)
-        if native.respond_to?(:instance_variable_defined?) &&
-            native.instance_variable_defined?(:@xml_decl)
-          # Explicitly set (programmatically added) - check if nil
-          !native.instance_variable_get(:@xml_decl).nil?
-        else
-          # Not set (parsed document) - trust wrapper flag
-          has_xml_declaration
-        end
-      when "Rexml"
-        # REXML: check @xml_declaration instance variable
-        # If not defined (parsed doc), trust wrapper flag
-        if native.respond_to?(:instance_variable_defined?) &&
-            native.instance_variable_defined?(:@xml_declaration)
-          # Explicitly set - check if nil
-          !native.instance_variable_get(:@xml_declaration).nil?
-        else
-          # Not set (parsed document) - trust wrapper flag
-          has_xml_declaration
-        end
-      when "Oga"
-        native.respond_to?(:xml_declaration) && !native.xml_declaration.nil?
-      when "Ox", "HeadedOx"
-        # Ox stores declaration in document attributes
-        native[:version] || native[:encoding] || native[:standalone]
-      when "Libxml"
-        # LibXML stores declaration wrapper as instance variable
-        if native.respond_to?(:instance_variable_defined?) &&
-            native.instance_variable_defined?(:@moxml_declaration)
-          # Explicitly set - check if nil
-          !native.instance_variable_get(:@moxml_declaration).nil?
-        else
-          # Not set - trust wrapper flag
-          has_xml_declaration
-        end
-      else
-        # Fallback - trust wrapper flag
-        has_xml_declaration
-      end
+      # For Document nodes, delegate to adapter for native state check
+      adapter.has_declaration?(@native, self)
     end
   end
 end

data/lib/moxml/node_set.rb

@@ -6,60 +6,72 @@ module Moxml
 
     attr_reader :nodes, :context
 
-    def initialize(nodes, context)
+    def initialize(nodes, context, parent_node = nil)
       @nodes = Array(nodes)
       @context = context
+      @wrapped = Array.new(@nodes.size)
+      @parent_node = parent_node
     end
 
     def each
       return to_enum(:each) unless block_given?
 
-      nodes.each { |node| yield Moxml::Node.wrap(node, context) }
+      @nodes.each_with_index do |node, i|
+        @wrapped[i] ||= wrap_with_parent(node)
+        yield @wrapped[i]
+      end
       self
     end
 
     def [](index)
       case index
       when Integer
-        Moxml::Node.wrap(nodes[index], context)
+        actual = index.negative? ? @nodes.size + index : index
+        return nil unless actual >= 0 && actual < @nodes.size
+
+        @wrapped[actual] ||= wrap_with_parent(@nodes[actual])
       when Range
-        NodeSet.new(nodes[index], context)
+        self.class.new(@nodes[index], @context)
       end
     end
 
     def first(n = nil)
       if n.nil?
-        Moxml::Node.wrap(nodes.first, context)
+        @nodes.empty? ? nil : self[0]
       else
-        nodes.first(n).map { |node| Moxml::Node.wrap(node, context) }
+        n.times.filter_map { |i| self[i] }
       end
     end
 
     def last
-      Moxml::Node.wrap(nodes.last, context)
+      @nodes.empty? ? nil : self[@nodes.size - 1]
     end
 
     def empty?
-      nodes.empty?
+      @nodes.empty?
     end
 
     def size
-      nodes.size
+      @nodes.size
     end
     alias length size
 
     def to_a
-      map { |node| node }
+      @nodes.each_with_index do |_node, i|
+        @wrapped[i] ||= wrap_with_parent(@nodes[i])
+      end
+      @wrapped.compact
     end
 
     def +(other)
-      self.class.new(nodes + other.nodes, context)
+      self.class.new(@nodes + other.nodes, @context)
     end
 
     def <<(node)
       # If it's a wrapped Moxml node, unwrap to native before storing
       native_node = node.respond_to?(:native) ? node.native : node
       @nodes << native_node
+      @wrapped << nil
       self
     end
     alias push <<
@@ -78,14 +90,14 @@ module Moxml
           true
         end
       end
-      self.class.new(unique_natives, context)
+      self.class.new(unique_natives, @context)
     end
 
     def ==(other)
       self.class == other.class &&
         length == other.length &&
-        nodes.each_with_index.all? do |node, index|
-          Moxml::Node.wrap(node, context) == other[index]
+        @nodes.each_with_index.all? do |_node, index|
+          self[index] == other[index]
         end
     end
 
@@ -103,8 +115,24 @@ module Moxml
     def delete(node)
       # If it's a wrapped Moxml node, unwrap to native
       native_node = node.respond_to?(:native) ? node.native : node
-      @nodes.delete(native_node)
+      idx = @nodes.index(native_node)
+      if idx
+        @nodes.delete_at(idx)
+        @wrapped.delete_at(idx)
+      else
+        @nodes.delete(native_node)
+      end
       self
     end
+
+    private
+
+    def wrap_with_parent(native_node)
+      wrapped = Moxml::Node.wrap(native_node, @context)
+      if @parent_node && wrapped
+        wrapped.parent_node = @parent_node
+      end
+      wrapped
+    end
   end
 end

data/lib/moxml/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Moxml
-  VERSION = "0.1.14"
+  VERSION = "0.1.16"
 end

data/lib/moxml/xml_utils.rb

@@ -81,7 +81,7 @@ module Moxml
     end
 
     def validate_prefix(prefix)
-      return if prefix.match?(/\A[a-zA-Z_][\w-]*\z/)
+      return if prefix.match?(/\A[a-zA-Z_][\w.-]*\z/)
 
       raise ValidationError, "Invalid namespace prefix: #{prefix}"
     end

data/lib/moxml/xpath/compiler.rb

@@ -1738,7 +1738,10 @@ module Moxml
         until visit.empty?
           current = visit.pop
 
-          return true if current.type == :call && current.children[0] == name
+          # Function name is stored in :value field, not children
+          if (current.type == :call || current.type == :function) && current.value == name
+            return true
+          end
 
           current.children.each do |child|
             visit << child if child.is_a?(AST::Node)

data/lib/moxml.rb

@@ -43,6 +43,7 @@ require_relative "moxml/builder"
 require_relative "moxml/config"
 require_relative "moxml/context"
 require_relative "moxml/entity_registry"
+require_relative "moxml/native_attachment"
 require_relative "moxml/adapter"
 require_relative "moxml/xpath"
 require_relative "moxml/sax"

data/scripts/format_xml.rb

@@ -0,0 +1,16 @@
+#!/usr/bin/env ruby
+# frozen_string_literal: true
+
+require "nokogiri"
+
+input = ARGV[0]
+output = ARGV[1] || input
+
+abort "Usage: #{$PROGRAM_NAME} <input.xml> [output.xml]" unless input
+abort "File not found: #{input}" unless File.exist?(input)
+
+doc = Nokogiri::XML(File.read(input), &:noblanks)
+formatted = doc.to_xml(indent: 2)
+
+File.write(output, formatted)
+puts "Formatted #{input}#{" -> #{output}" if output != input}"

data/scripts/pretty_format_xml.rb

@@ -0,0 +1,14 @@
+#!/usr/bin/env ruby
+# frozen_string_literal: true
+
+require "nokogiri"
+
+input = ARGV[0] or abort "Usage: #{$0} <input.xml> [output.xml]"
+output = ARGV[1] || input
+
+xml = File.read(input)
+doc = Nokogiri::XML(xml, &:noblanks)
+formatted = doc.to_xml(indent: 2)
+
+File.write(output, formatted)
+puts "Written to #{output}"

data/spec/consistency/round_trip_spec.rb

@@ -62,14 +62,9 @@ def traverse_with_consistent_order(element, elements_array)
   end
 
   if element.respond_to?(:children)
-    # ENHANCED: More robust child selection and sorting
+    # Only traverse element children (skip text, comment, cdata nodes)
     children = element.children.select do |child|
-      # Only process element nodes with valid names
-      child.respond_to?(:name) &&
-        child.name &&
-        !child.name.empty? &&
-        child.name != "text" &&
-        child.name != "comment"
+      child.respond_to?(:element?) && child.element?
     end
 
     # CRITICAL: Enhanced sorting with multiple criteria for stability
@@ -324,29 +319,7 @@ EXAMPLE_TIMEOUT = ENV.fetch("MOXML_ROUNDTRIP_TIMEOUT", 120).to_i
 # Fixture cache — loaded once, shared across all examples.
 FIXTURE_CACHE = {}
 
-# Known element ordering issues with Ox adapter.
-# These (fixture_relative_path, source_adapter, target_adapter) tuples fail the
-# elements_with_attributes comparison because Ox produces elements in a different
-# order. The semantic equivalence check (double round-trip) still passes.
-# TODO: Investigate and fix the root cause in ox adapter element ordering.
-KNOWN_ELEMENT_ORDERING_ISSUES = Set.new([
-  # niso-jats/element_citation.xml - Ox produces different element ordering
-  ["niso-jats/element_citation.xml", :nokogiri, :ox],
-  ["niso-jats/element_citation.xml", :ox, :nokogiri],
-  ["niso-jats/element_citation.xml", :ox, :oga],
-  ["niso-jats/element_citation.xml", :oga, :ox],
-  ["niso-jats/element_citation.xml", :rexml, :ox],
-  ["niso-jats/element_citation.xml", :ox, :rexml],
-  ["niso-jats/pnas_sample.xml", :nokogiri, :rexml],
-  ["niso-jats/pnas_sample.xml", :rexml, :nokogiri],
-  # metanorma fixtures with similar issues
-  ["metanorma/collection1nested.xml", :nokogiri, :ox],
-  ["metanorma/collection1nested.xml", :ox, :nokogiri],
-  ["metanorma/collection1nested.xml", :ox, :oga],
-  ["metanorma/collection1nested.xml", :oga, :ox],
-  ["metanorma/collection1nested.xml", :rexml, :ox],
-  ["metanorma/collection1nested.xml", :ox, :rexml],
-])
+KNOWN_ELEMENT_ORDERING_ISSUES = Set.new([])
 
 RSpec.describe "Round-trip XML Testing", :round_trip do
   # Explicit adapter names for clarity and maintainability.

data/spec/integration/all_adapters_spec.rb

@@ -15,6 +15,7 @@ RSpec.describe "Cross-adapter integration" do
     "Moxml::Declaration",
     "Moxml::Doctype",
     "Moxml::Document",
+    "Moxml::EntityReference",
     "Moxml::Context",
     "Moxml::Builder",
     "Moxml::DocumentBuilder",

data/spec/integration/headed_ox_integration_spec.rb

@@ -187,7 +187,6 @@ RSpec.describe "HeadedOx Integration" do
       end
 
       it "selects last node with last()" do
-        skip "HeadedOx limitation: last() in predicate context needs temporary nodeset. See docs/HEADED_OX_LIMITATIONS.md"
         last_book = doc.xpath("//book[position() = last()]")
 
         expect(last_book.size).to eq(1)
@@ -198,7 +197,6 @@ RSpec.describe "HeadedOx Integration" do
 
   describe "real-world use cases" do
     it "finds books by author and price range" do
-      skip "HeadedOx limitation: Text content access from nested elements needs investigation. See docs/HEADED_OX_LIMITATIONS.md"
       results = doc.xpath(
         "//book[contains(author, 'Alice') and @price < 20]",
       )

data/spec/integration/shared_examples/edge_cases.rb

@@ -36,7 +36,7 @@ RSpec.shared_examples "Moxml Edge Cases" do
         pending "Ox doesn't escape the end token"
       end
       if context.config.adapter_name == :headed_ox
-        skip "HeadedOx limitation: Ox doesn't escape CDATA end markers. See docs/HEADED_OX_LIMITATIONS.md"
+        skip "HeadedOx limitation: Ox doesn't escape CDATA end markers. See docs/_pages/headed-ox-limitations.adoc"
       end
       cdata_text = "]]>]]>]]>"
       doc = context.create_document
@@ -90,11 +90,14 @@ RSpec.shared_examples "Moxml Edge Cases" do
         pending "Ox doesn't have a native XPath"
       end
       if context.config.adapter_name == :headed_ox
-        skip "HeadedOx limitation: Namespace methods not implemented in adapter. Requires Ox namespace API enhancement. See docs/HEADED_OX_LIMITATIONS.md"
+        skip "HeadedOx limitation: Namespace methods not implemented in adapter. Requires Ox namespace API enhancement. See docs/_pages/headed-ox-limitations.adoc"
       end
       if context.config.adapter_name == :libxml
         skip "LibXML cannot query empty default namespace with XPath (documented limitation)"
       end
+      if context.config.adapter_name == :nokogiri
+        skip "Nokogiri XPath does not support querying empty namespace with xmlns prefix mapping"
+      end
       xml = <<~XML
         <root xmlns="http://default1.org">
           <child xmlns="http://default2.org">
@@ -113,7 +116,7 @@ RSpec.shared_examples "Moxml Edge Cases" do
         pending "Ox doesn't have a native XPath"
       end
       if context.config.adapter_name == :headed_ox
-        skip "HeadedOx limitation: Namespace methods not implemented in adapter. Requires Ox namespace API enhancement. See docs/HEADED_OX_LIMITATIONS.md"
+        skip "HeadedOx limitation: Namespace methods not implemented in adapter. Requires Ox namespace API enhancement. See docs/_pages/headed-ox-limitations.adoc"
       end
 
       xml = <<~XML
@@ -133,7 +136,7 @@ RSpec.shared_examples "Moxml Edge Cases" do
   describe "attribute edge cases" do
     it "handles attributes with same local name but different namespaces" do
       if context.config.adapter_name == :headed_ox
-        skip "HeadedOx limitation: Namespace-prefixed attribute access needs Ox namespace API enhancement. See docs/HEADED_OX_LIMITATIONS.md"
+        skip "HeadedOx limitation: Namespace-prefixed attribute access needs Ox namespace API enhancement. See docs/_pages/headed-ox-limitations.adoc"
       end
       if context.config.adapter_name == :ox
         skip "Ox doesn't have a native XPath"

data/spec/integration/shared_examples/integration_workflows.rb

@@ -60,7 +60,7 @@ RSpec.shared_examples "Moxml Integration" do
         pending "Ox doesn't support namespace-aware XPath with predicates"
       end
       if context.config.adapter_name == :headed_ox
-        skip "HeadedOx limitation: Namespace-aware XPath with predicates needs investigation. See docs/HEADED_OX_LIMITATIONS.md"
+        skip "HeadedOx limitation: Namespace-aware XPath with predicates needs investigation. See docs/_pages/headed-ox-limitations.adoc"
       end
       # Test XPath queries
       #
@@ -80,7 +80,7 @@ RSpec.shared_examples "Moxml Integration" do
         pending "Ox doesn't have a native XPath"
       end
       if context.config.adapter_name == :headed_ox
-        skip "HeadedOx limitation: Namespace methods not implemented in adapter. Requires Ox namespace API enhancement. See docs/HEADED_OX_LIMITATIONS.md"
+        skip "HeadedOx limitation: Namespace methods not implemented in adapter. Requires Ox namespace API enhancement. See docs/_pages/headed-ox-limitations.adoc"
       end
       xml = <<~XML
         <root xmlns="http://default.org" xmlns:a="http://a.org" xmlns:b="http://b.org">
@@ -123,7 +123,7 @@ RSpec.shared_examples "Moxml Integration" do
 
     it "handles complex modifications" do
       if context.config.adapter_name == :headed_ox
-        skip "HeadedOx limitation: Parent setter not implemented. Requires Ox node reparenting API. See docs/HEADED_OX_LIMITATIONS.md"
+        skip "HeadedOx limitation: Parent setter not implemented. Requires Ox node reparenting API. See docs/_pages/headed-ox-limitations.adoc"
       end
       if context.config.adapter_name == :ox
         skip "Ox doesn't have a native XPath"

data/spec/integration/shared_examples/node_wrappers/cdata_behavior.rb

@@ -41,7 +41,7 @@ RSpec.shared_examples "Moxml::Cdata" do
         pending "Ox doesn't escape the end token"
       end
       if context.config.adapter_name == :headed_ox
-        skip "HeadedOx limitation: Ox doesn't escape CDATA end markers. See docs/HEADED_OX_LIMITATIONS.md"
+        skip "HeadedOx limitation: Ox doesn't escape CDATA end markers. See docs/_pages/headed-ox-limitations.adoc"
       end
       cdata.content = "content]]>more"
       expect(cdata.to_xml).to eq("<![CDATA[content]]]]><![CDATA[>more]]>")