canon 0.3.98 → 0.3.99

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 07c9d8c96426b670ad088e3b96e662b45283c4868e99c27a5afb5a6ecb537780
-  data.tar.gz: 5e8bc5ee7aa5b1e1b75b158d1afc68f3035e1c58bcd3da4624fdebbbbc83b4a6
+  metadata.gz: 4e993fe0bac445ec82e1a02983bfa2f297c1b7d9d112bcb7c724f3f010a1b353
+  data.tar.gz: 869324e74378af4f4cb9493fbfdc04ac1ba8e76c0378159547bc7e23ebf66084
 SHA512:
-  metadata.gz: 813b81bd8fe090e71c9729d18326e864b0af54129e422ae4d624676ed8e191f6b2b278e5dab5e00c1827fccaed9658e5bd85a9ef6ac908c9e0cc4ff97d33839e
-  data.tar.gz: 6819b9252b0e6f27efec816241b1444d1ff92456562996ddc0c89a8f07b04eb41d48302025fda8776581437fb793da24db3ed21fec2b1589522c0680739bd99c
+  metadata.gz: 70c6dcb6bb9da7ef764f9aed9e1e8d02fb95193c446102df62496a10dbe33010d29b5d0e90293d929cf3adc267b5128ad1dd7e9666dff38b892d221bc681f20c
+  data.tar.gz: dc45835bdab2c58479e8413eb87aee5c8615237ebf6bd2c435ae725b58b7df247c3740c25a3ab192fb0aaf86c6776d3634c5a66170ea6834a3f3bedf5582fd27

data/lib/canon/comparison/html_comparator.rb

@@ -295,8 +295,8 @@ module Canon
                              Canon::Xml::C14n.canonicalize(html_string,
                                                            with_comments: false)
                            when :format
-                             # Pretty format
-                             Canon.format(html_string, :html)
+                             # Engine single-line structural normalizer
+                             Canon::Html::DataModel.engine_html_format(html_string)
                            else
                              # :none or unrecognized
                              html_string
@@ -375,8 +375,8 @@ module Canon
                           Canon::Xml::C14n.canonicalize(node,
                                                         with_comments: false)
                         when :format
-                          # Pretty format the HTML
-                          Canon.format(node, :html)
+                          # Engine single-line structural normalizer
+                          Canon::Html::DataModel.engine_html_format(node)
                         else
                           # :none, :rendered or unrecognized - use as-is
                           node

data/lib/canon/comparison/html_parser.rb

@@ -60,29 +60,10 @@ module Canon
         end
 
         # Parse an HTML string through the engine's WHATWG fragment
-        # parsing and build a Canon tree from the fragment children.
-        # Whitespace text nodes are preserved unconditionally — HTML
-        # whitespace significance is the comparison profile's decision
-        # (#112/#137 ride on stray whitespace surviving the parse),
-        # unlike the XML DOM lane's parse-time policy. Top-level
-        # element children convert through the materialize-records
-        # fast path; the wrapper walk serves the rare non-element
-        # top-level nodes (text, comments).
+        # parsing and build a Canon tree. Canon::Html::DataModel owns
+        # the conversion (the single home for HTML→Canon-tree building).
         def engine_fragment(html)
-          frag = Canon::XmlParsing.moxml_context.parse_html_fragment(html)
-          root = Canon::Xml::Nodes::RootNode.new
-          frag.children.each do |child|
-            node =
-              if child.is_a?(Moxml::Element)
-                Canon::Xml::DataModel.build_moxml_subtree_from_records(
-                  child, preserve_whitespace: true
-                )
-              else
-                Canon::Xml::DataModel.walk_moxml(child, preserve_whitespace: true)
-              end
-            root.add_child(node) if node
-          end
-          root
+          Canon::Html::DataModel.engine_fragment(html)
         end
 
         # Engine HTML fragments are served by adapters with an engine

data/lib/canon/html/data_model.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-Canon::NokogiriLoader.require!("HTML support")
-
 module Canon
   module Html
     # Builds XPath data model from HTML
@@ -10,58 +8,53 @@ module Canon
     class DataModel < Canon::DataModel
       # Build XPath data model from HTML string
       #
+      # The engine lane: moxml Context#parse_html_fragment (WHATWG
+      # fragment parsing) → Canon tree. The +version:+ parameter is
+      # accepted for API compatibility — the engine serves every input
+      # through one WHATWG parse (the #118 html4/html5 unification).
+      # Full documents keep their head content at the fragment top
+      # level (the in-body rules ignore the html/body start tags but
+      # preserve every other node); the former Nokogiri body-extraction
+      # dropped head content entirely.
+      #
       # @param html_string [String] HTML content to parse
-      # @param version [Symbol] HTML version (:html4 or :html5)
+      # @param version [Symbol] accepted for compatibility; unused
       # @return [Canon::Xml::Nodes::RootNode] Root of the data model tree
+      # rubocop:disable-next Lint/UnusedMethodArgument -- API compatibility
       def self.from_html(html_string, version: :html4)
-        # Detect if this is a full document (has <html> tag) or fragment
-        # Full documents should use document parser to preserve structure
-        # Fragments should use fragment parser to avoid adding implicit wrappers
-        is_full_document = html_string.match?(%r{<html[\s>]}i)
-
-        # Parse with Nokogiri using appropriate parser
-        doc = if is_full_document
-                # CRITICAL FIX: For full HTML documents, parse as document first
-                # and extract the body element. This avoids Nokogiri::HTML.fragment()
-                # incorrectly moving head elements (like meta) to the body.
-                # Parse as full document to get proper structure
-                full_doc = if version == :html5
-                             Nokogiri::HTML5(html_string)
-                           else
-                             Nokogiri::HTML4(html_string)
-                           end
-                # Extract body element and create fragment from it
-                body = full_doc.at_css("body")
-                if body
-                  # Create a fragment and copy body children to it
-                  # This preserves the body structure without head elements
-                  frag = if version == :html5
-                           Nokogiri::HTML5::DocumentFragment.new(full_doc)
-                         else
-                           Nokogiri::HTML4::DocumentFragment.new(full_doc)
-                         end
-                  body.children.each do |child|
-                    frag.add_child(child.dup)
-                  end
-                  frag
-                elsif version == :html5
-                  # No body found, fall back to fragment parsing
-                  Nokogiri::HTML5.fragment(html_string)
-                else
-                  Nokogiri::HTML4.fragment(html_string)
-                end
-              elsif version == :html5
-                # Fragment - use fragment parser to avoid implicit wrappers
-                Nokogiri::HTML5.fragment(html_string)
-              else
-                Nokogiri::HTML4.fragment(html_string)
-              end
+        engine_fragment(html_string)
+      end
 
-        # HTML doesn't have strict namespace requirements like XML,
-        # so skip the relative namespace URI check
+      # Parse an HTML string through the engine's WHATWG fragment
+      # parsing and build a Canon tree from the fragment children.
+      # Whitespace text nodes are preserved unconditionally — HTML
+      # whitespace significance is the comparison profile's decision
+      # (#112/#137), unlike the XML DOM lane's parse-time policy.
+      # Top-level element children convert through the
+      # materialize-records fast path; the wrapper walk serves the
+      # rare non-element top-level nodes (text, comments).
+      def self.engine_fragment(html)
+        frag = Canon::XmlParsing.moxml_context.parse_html_fragment(html)
+        root = Canon::Xml::Nodes::RootNode.new
+        frag.children.each do |child|
+          node =
+            if child.is_a?(Moxml::Element)
+              Canon::Xml::DataModel.build_moxml_subtree_from_records(
+                child, preserve_whitespace: true
+              )
+            else
+              Canon::Xml::DataModel.walk_moxml(child, preserve_whitespace: true)
+            end
+          root.add_child(node) if node
+        end
+        root
+      end
 
-        # Convert to XPath data model (reuse XML infrastructure)
-        build_from_nokogiri(doc)
+      # Deterministic single-line engine formatting for the :format
+      # string preprocessor — parse + re-serialize as a pure function
+      # of the parsed tree.
+      def self.engine_html_format(html)
+        Canon::XmlParsing.moxml_context.parse_html(html).to_html
       end
 
       # Alias for compatibility

data/lib/canon/tree_diff/adapters/html_adapter.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-Canon::NokogiriLoader.require!("HTML tree diffing")
-
 module Canon
   module TreeDiff
     module Adapters
@@ -46,17 +44,22 @@ module Canon
             return to_tree_from_canon_comment(node)
           end
 
-          # Fallback to Nokogiri (legacy support)
-          case node
-          when Nokogiri::HTML::Document, Nokogiri::HTML4::Document, Nokogiri::HTML5::Document
-            # Start from html element or root element
-            root = node.at_css("html") || node.root
-            root ? to_tree(root) : nil
-          when Nokogiri::HTML4::DocumentFragment, Nokogiri::HTML5::DocumentFragment, Nokogiri::XML::DocumentFragment
-            # For DocumentFragment, create a wrapper root node and add all fragment children
-            convert_fragment(node)
-          when Nokogiri::XML::Element
-            convert_element(node)
+          # Fallback to Nokogiri (legacy support for user-supplied
+          # nodes) — the Canon tree branches above run nokogiri-less.
+          if defined?(Nokogiri)
+            case node
+            when Nokogiri::HTML::Document, Nokogiri::HTML4::Document, Nokogiri::HTML5::Document
+              # Start from html element or root element
+              root = node.at_css("html") || node.root
+              root ? to_tree(root) : nil
+            when Nokogiri::HTML4::DocumentFragment, Nokogiri::HTML5::DocumentFragment, Nokogiri::XML::DocumentFragment
+              # For DocumentFragment, create a wrapper root node and add all fragment children
+              convert_fragment(node)
+            when Nokogiri::XML::Element
+              convert_element(node)
+            else
+              raise ArgumentError, "Unsupported node type: #{node.class}"
+            end
           else
             raise ArgumentError, "Unsupported node type: #{node.class}"
           end
@@ -68,7 +71,10 @@ module Canon
         # @param doc [Nokogiri::HTML::Document] Optional document to use
         # @return [Nokogiri::HTML::Document, Nokogiri::XML::Element]
         def from_tree(tree_node, doc = nil)
-          doc ||= Nokogiri::HTML::Document.new
+          unless doc
+            Canon::NokogiriLoader.require!("HTML tree reconstruction")
+            doc = Nokogiri::HTML::Document.new
+          end
 
           element = build_element(tree_node, doc)
 

data/lib/canon/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Canon
-  VERSION = "0.3.98"
+  VERSION = "0.3.99"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: canon
 version: !ruby/object:Gem::Version
-  version: 0.3.98
+  version: 0.3.99
 platform: ruby
 authors:
 - Ribose Inc.