canon 0.3.108 → 0.3.110

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 58b0a315d811eb36769d72ef0949180ae5c6eb3b23e4732887783d36483dc8f8
-  data.tar.gz: 739ec6e878fb031f6ca8814587f3763952422679858a992d66b45a541e85600f
+  metadata.gz: 831a2d23d844114132e9a234a4e1452f08aacc36ef0947d80c7100358bbbe209
+  data.tar.gz: ff423e64f21b9e5e6c93694c124fc404b95e31ad5b4ed3162262c62023bd5c35
 SHA512:
-  metadata.gz: 0a57c1c8062483a762edec9caaef0ee8432f03f88b518163ab9b7836b52c6c671025bf9db6525e870362976c380e1fe0ba7487eb1447e4cdaf2f88538744e89d
-  data.tar.gz: 17a52ab1a74f3d34d34fcf9f36fd5c76104be437e5e9bd16d74500306d539863a0233048ddde79078be6fb993e25b3530287006cf735007945200cef29acdaeb
+  metadata.gz: 35762c78b2ba5bd8090cb799566ecabec4d863bd956c27c0a9119bf8025cda2bad8d49f9a93b4242ece7bfc5de7e8ef5a939bce5e3f955e46af925857379e704
+  data.tar.gz: 4b932b9b03940bc536887cbf4a750ce265b2992afc775b446ea9a56b675d23f52724b7b7a6f710ae3258a4c5e73c69bfc01bd67b56087554ca0a3c47a5f35343

data/lib/canon/comparison/xml_comparator/child_comparison.rb

@@ -198,6 +198,12 @@ module Canon
           )
             has_mismatch = false
 
+            # The ORIGINAL direction, captured before
+            # determine_mismatch_children returns reduced sets of equal
+            # length — deriving the side from the reduced sets reports
+            # every mid-content removal as an insertion (canon#285).
+            before_longer = children1.length > children2.length
+
             # Length check
             unless children1.length == children2.length
               has_mismatch = true
@@ -228,7 +234,7 @@ module Canon
               else
                 mismatched_children.each do |child|
                   child_dim = comparator.determine_node_dimension(child)
-                  if children1.length > children2.length
+                  if before_longer
                     comparator.add_difference(child, nil,
                                               Comparison::MISSING_NODE,
                                               Comparison::MISSING_NODE,

data/lib/canon/formatters/html_formatter_base.rb

@@ -63,6 +63,12 @@ module Canon
       # regex per format call cost more than the gsub it drives.
       BLOCK_SPACING_PATTERN =
         Regexp.new("(</(?:#{BLOCK_ELEMENTS.join('|')})>)(?:\n*)(<(?:#{BLOCK_ELEMENTS.join('|')})[\s>])").freeze
+      # Engine-inserted newlines after a block close, before ANY next
+      # tag (the serializer's indent model breaks lines where canon's
+      # normalize step kept no text node) — folded to the canonical
+      # single space so serialization is idempotent.
+      AFTER_BLOCK_NEWLINE_PATTERN =
+        Regexp.new("(</(?:#{BLOCK_ELEMENTS.join('|')})>)(\n+)(<(?!/))").freeze
       # Format HTML using canonical form
       # @param html [String] HTML document to canonicalize
       # @return [String] Canonical form of HTML
@@ -108,21 +114,10 @@ module Canon
       end
 
       # Sort element attributes alphabetically throughout document
+      # (the shared canonical attribute order — TODO.sota/03)
       # @param doc [Moxml::Document] Parsed document
       def self.sort_attributes(doc)
-        walk_elements(doc.root) do |element|
-          attrs = element.attributes
-          next if attrs.empty?
-
-          names = attrs.map(&:name)
-          # Already-sorted is the common case — removing and re-adding
-          # every attribute is expensive, so check first.
-          next if names.each_cons(2).all? { |a, b| (a <=> b) <= 0 }
-
-          sorted_attrs = attrs.sort_by(&:name)
-          names.each { |name| element.remove_attribute(name) }
-          sorted_attrs.each { |attr| element[attr.name] = attr.value }
-        end
+        Canon::Xml::AttributeSorter.sort!(doc)
       end
 
       # Depth-first element walk (pre-order).
@@ -211,9 +206,17 @@ module Canon
       # @param html [String] Serialized HTML string
       # @return [String] HTML with proper spacing between block elements
       def self.ensure_block_element_spacing(html)
-        # Add space between closing and opening block element tags
+        # Space between closing and opening block element tags
         # (pattern compiled once — see BLOCK_SPACING_PATTERN)
-        html.gsub(BLOCK_SPACING_PATTERN, '\1 \2')
+        html = html.gsub(BLOCK_SPACING_PATTERN, '\1 \2')
+        # The engine's serializer inserts its own newlines after a
+        # block close when the tree has no whitespace text node there
+        # (normalize_whitespace keeps a single space only where one
+        # existed). Collapse those self-inserted newlines to the same
+        # canonical space — without this, format is not idempotent:
+        # pass 1 emits "</p>\n<br>", pass 2 answers "</p> <br>"
+        # (the property suite, TODO.sota/03).
+        html.gsub(AFTER_BLOCK_NEWLINE_PATTERN, '\1 \3')
       end
 
       # Check if a node is a block-level element

data/lib/canon/formatters/xml_formatter.rb

@@ -11,13 +11,9 @@ module Canon
         end
       end
 
-      def self.parse(xml)
+      def self.parse(xml, options = {})
         Canon::Validators::XmlValidator.validate!(xml)
-        if XmlBackend.nokogiri?
-          Nokogiri::XML(xml)
-        else
-          XmlParsing.parse(xml)
-        end
+        XmlParsing.parse(xml, options)
       end
     end
   end

data/lib/canon/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Canon
-  VERSION = "0.3.108"
+  VERSION = "0.3.110"
 end

data/lib/canon/xml/attribute_sorter.rb

@@ -0,0 +1,39 @@
+# frozen_string_literal: true
+
+module Canon
+  module Xml
+    # Canonical attribute ordering: every element's attributes are
+    # re-attached in lexicographic name order. Both canonical lanes
+    # consume this — the HTML formatter base and the XML pretty
+    # printer — so attribute order is invariant across formats and
+    # across input orderings (the equivalence-respect property,
+    # TODO.sota/03).
+    module AttributeSorter
+      module_function
+
+      def sort!(doc)
+        walk_elements(doc.root) do |element|
+          attrs = element.attributes
+          next if attrs.empty?
+
+          names = attrs.map(&:name)
+          # Already-sorted is the common case — removing and re-adding
+          # every attribute is expensive, so check first.
+          next if names.each_cons(2).all? { |a, b| (a <=> b) <= 0 }
+
+          sorted_attrs = attrs.sort_by(&:name)
+          names.each { |name| element.remove_attribute(name) }
+          sorted_attrs.each { |attr| element[attr.name] = attr.value }
+        end
+        doc
+      end
+
+      def walk_elements(node, &block)
+        return unless node.is_a?(Moxml::Element)
+
+        yield(node)
+        node.children.to_a.each { |child| walk_elements(child, &block) }
+      end
+    end
+  end
+end

data/lib/canon/xml.rb

@@ -15,6 +15,7 @@ module Canon
   # Canon::Xml::Nodes.
   module Xml
     autoload :AttributeHandler, "canon/xml/attribute_handler"
+    autoload :AttributeSorter, "canon/xml/attribute_sorter"
     autoload :C14n, "canon/xml/c14n"
     autoload :DigestGate, "canon/xml/digest_gate"
     autoload :CharacterEncoder, "canon/xml/character_encoder"

data/lib/canon/xml_parsing.rb

@@ -303,11 +303,14 @@ module Canon
 
       # --- Moxml engine ---
 
-      # Readonly: canon only ever reads engine documents (conversion,
+      # Readonly by default: canon reads engine documents (conversion,
       # serialization) — leptris memoizes reads on readonly documents,
       # and mutation is refused loudly rather than silently corrupting.
-      def moxml_parse(xml_string, _options)
-        moxml_context.parse(xml_string, readonly: true)
+      # Mutating lanes (the attribute sorter, TODO.sota/03) opt out
+      # explicitly with readonly: false.
+      def moxml_parse(xml_string, options)
+        readonly = options.fetch(:readonly, true)
+        moxml_context.parse(xml_string, readonly: readonly)
       end
 
       def moxml_node_type(node)

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: canon
 version: !ruby/object:Gem::Version
-  version: 0.3.108
+  version: 0.3.110
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-10-21 00:00:00.000000000 Z
+date: 2026-10-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: diff-lcs
@@ -405,6 +405,7 @@ files:
 - lib/canon/version.rb
 - lib/canon/xml.rb
 - lib/canon/xml/attribute_handler.rb
+- lib/canon/xml/attribute_sorter.rb
 - lib/canon/xml/c14n.rb
 - lib/canon/xml/character_encoder.rb
 - lib/canon/xml/data_model.rb