canon 0.3.59 → 0.3.61

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 59a766120de370a517050c389b3ec809574e8e0f876b6a9b3ae351386cfc7333
-  data.tar.gz: 6a64cdc990d0d295645ec1f893e37da8b03858c06a2b3992d0e00926d04c939b
+  metadata.gz: b5b90b7fc1d45d8f8e58e804f3f67dae934cc225639f3196c64f43cdd84ae005
+  data.tar.gz: c6615312bb2ac6eba00700c60b7dfeeb18277b8677532f98f91e4ad24a29c562
 SHA512:
-  metadata.gz: 0f659e6d2708dc675e2bc0df68eb57773ca16224083249e95a9906b1f7d967c92a917882f73a810490ca3b11c228afb9b581d1ceb80b0dc674d98d46014aba42
-  data.tar.gz: 974a7dfa4cdcacb21e76952c193160022ecd83de6887b4a5582c60d5853feb5c89a3e69be559ef36484c37a8f9228da1cad8a9168829d23b3e096d51d09fca6d
+  metadata.gz: 75e2773c9597d616988157bdd40f4df3478dfd5695c1b9c7b9fc3709a6b043e22e98d3a1d9f2221ad06f7d4bb249331b1356606dce4acf10ce4bdfd8c254d209
+  data.tar.gz: 2877cc2367ac8da5ef68ea9b2ba2cc94883754a47aaf7dba51aef1050d67b9f3d95027238a539e119c92704fa696bdd8de8412bf178098a00e03061fbbb48888

data/lib/canon/comparison/xml_comparator.rb

@@ -201,11 +201,17 @@ module Canon
         # side answers both; any hit falls through to the pipeline.
         def verbose_report_proven_empty?(n1, n2, opts)
           return true unless opts[:verbose]
+
           # Identical input parses identically — differences and
-          # attribute order coincide by construction; only the SAX
-          # recover-error surface (parse-error banner, issue #130)
-          # needs its one scan.
-          return !Xml::Sax.probe(n1).saw_error? if n1 == n2
+          # attribute order coincide by construction. Recover errors
+          # ride the fingerprint's diag check when the engine exposes
+          # them (libleptris 1.9.205+, #1200); otherwise the one SAX
+          # scan covers the parse-error banner (issue #130).
+          if n1 == n2
+            return true if Xml::DigestGate.recover_diags_available?
+
+            return !Xml::Sax.probe(n1).saw_error?
+          end
 
           left = Xml::Sax.probe(n1)
           return false if left.saw_error?

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

@@ -196,7 +196,7 @@ module Canon
           return false unless Canon::Comparison::NodeInspector.element_node?(element)
 
           # List of HTML elements where whitespace is semantically significant
-          whitespace_sensitive_tags = %w[pre code textarea script style]
+          whitespace_sensitive_tags = Core::TreeNode::WHITESPACE_SENSITIVE_TAGS
           whitespace_sensitive_tags.include?(element.name.downcase)
         end
 

data/lib/canon/tree_diff/core/tree_node.rb

@@ -87,8 +87,9 @@ module Canon
           children.empty?
         end
 
-        # Elements where whitespace is semantically significant; their
-        # text value participates in the signature (see NodeSignature).
+        # Elements where whitespace is semantically significant (HTML):
+        # their text value participates in signatures and matching.
+        # Single source for the tree_diff components.
         WHITESPACE_SENSITIVE_TAGS = %w[pre code textarea script style].freeze
 
         def build_signature_path(include_attributes)
@@ -109,9 +110,17 @@ module Canon
           component = label.to_s
 
           # Sorted attributes distinguish same-label nodes while
-          # ignoring attribute order.
+          # ignoring attribute order. Built directly into the
+          # component — no intermediate sorted/map/join arrays.
           if include_attributes && !attributes.empty?
-            component += "{#{attributes.sort.map { |k, v| "#{k}=#{v}" }.join(',')}}" # rubocop:disable Style/StringConcatenation
+            component += "{"
+            first = true
+            attributes.sort.each do |k, v|
+              component << "," unless first
+              first = false
+              component << k.to_s << "=" << v.to_s
+            end
+            component << "}"
           end
 
           if include_attributes &&
@@ -165,10 +174,14 @@ module Canon
         #
         # @return [Array<TreeNode>]
         def descendants
+          # Iterative pre-order walk — the recursive form allocated an
+          # intermediate array per level.
           result = []
-          children.each do |child|
-            result << child
-            result.concat(child.descendants)
+          stack = children.reverse_each.to_a
+          until stack.empty?
+            node = stack.pop
+            result << node
+            stack.concat(node.children.reverse_each.to_a) if node.children.any?
           end
           result
         end

data/lib/canon/tree_diff/matchers/similarity_matcher.rb

@@ -158,7 +158,7 @@ module Canon
           return false unless node
 
           # List of HTML elements where whitespace is semantically significant
-          whitespace_sensitive_tags = %w[pre code textarea script style]
+          whitespace_sensitive_tags = Core::TreeNode::WHITESPACE_SENSITIVE_TAGS
 
           # Check if this node is whitespace-sensitive
           if node.is_a?(Core::TreeNode)

data/lib/canon/tree_diff/matchers/structural_propagator.rb

@@ -211,7 +211,7 @@ module Canon
           return false unless node
 
           # List of HTML elements where whitespace is semantically significant
-          whitespace_sensitive_tags = %w[pre code textarea script style]
+          whitespace_sensitive_tags = Core::TreeNode::WHITESPACE_SENSITIVE_TAGS
 
           # Check if this node is whitespace-sensitive
           if node.is_a?(Core::TreeNode)

data/lib/canon/tree_diff/operations/operation_detector.rb

@@ -205,8 +205,10 @@ module Canon
           attrs1 = node1.attributes
           attrs2 = node2.attributes
 
-          # Check if attribute values differ (ignoring order)
-          if attrs1.sort.to_h != attrs2.sort.to_h
+          # Check if attribute values differ (ignoring order).
+          # Identical raw hashes (order included) are trivially equal
+          # after sorting — the overwhelmingly common matched case.
+          if attrs1 != attrs2 && attrs1.sort.to_h != attrs2.sort.to_h
             # Actual attribute value differences
             changes[:attributes] = {
               old: attrs1,
@@ -611,15 +613,12 @@ module Canon
         def whitespace_sensitive?(node)
           return false unless node
 
-          # List of HTML elements where whitespace is semantically significant
-          whitespace_sensitive_tags = %w[pre code textarea script style]
-
           # Check if this node or any ancestor is whitespace-sensitive
           current = node
           while current
             if current.is_a?(Core::TreeNode)
               label = current.label.to_s.downcase
-              return true if whitespace_sensitive_tags.include?(label)
+              return true if Core::TreeNode::WHITESPACE_SENSITIVE_TAGS.include?(label)
             end
 
             current = current.is_a?(Core::TreeNode) ? current.parent : nil

data/lib/canon/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Canon
-  VERSION = "0.3.59"
+  VERSION = "0.3.61"
 end

data/lib/canon/xml/digest_gate.rb

@@ -21,6 +21,28 @@ module Canon
     module DigestGate
       module_function
 
+      # libleptris 1.9.205+ records duplicate-attribute recover
+      # events on the parsed document (#1200): read through the
+      # FFI diag counter. Evaluated lazily at call time — a load-time
+      # constant would freeze before leptris loads (defined? never
+      # resolves pending autoloads). When the adapter is not leptris
+      # (or the binding predates the surface), the verbose lane
+      # falls back to the SAX probe for its error scan.
+      def recover_diags_available?
+        defined?(::Leptris::XML::FFI) &&
+          ::Leptris::XML::FFI.respond_to?(:leptris_document_parse_diag_count)
+      rescue StandardError
+        false
+      end
+
+      def recover_diags?(doc)
+        return false unless recover_diags_available?
+
+        native = doc.native
+        native = native.c_ptr if native.respond_to?(:c_ptr)
+        ::Leptris::XML::FFI.leptris_document_parse_diag_count(native).positive?
+      end
+
       def available?
         return false if RUBY_ENGINE == "opal"
         return false unless Canon::XmlBackend.moxml? &&
@@ -106,7 +128,11 @@ module Canon
         doc = context.parse(xml, readonly: true, strict: false)
         root = doc.root
         return nil unless root
-        return nil if clean && doc.parse_errors.any?
+
+        if clean
+          return nil if doc.parse_errors.any?
+          return nil if recover_diags?(doc)
+        end
 
         skeleton = doc.children.filter_map do |child|
           next if Canon::XmlParsing.same_engine_node?(child, root)

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: canon
 version: !ruby/object:Gem::Version
-  version: 0.3.59
+  version: 0.3.61
 platform: ruby
 authors:
 - Ribose Inc.