canon 0.3.16 → 0.3.18

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 133da0e86ad37cf095f738d15a37ebcc3b35ae92240327d459ed1d1bbedc3a5b
-  data.tar.gz: 6c7c2cf6e60712058a6f49f28cf9181316f9e410da5a3982b72fee7fdebde115
+  metadata.gz: b27c3659e1b75307f170fb50161ce649c54ab4a12600716d3b006fce8b01ad1a
+  data.tar.gz: 89c288d724f4a3b0b410938d94959880b4fbbcff1ec550b53593840bd33d52c8
 SHA512:
-  metadata.gz: fd39028ce4e945d04ac83d44f9a0da8b516dadf8a454bb634147aabfa480f896ed25f7cf048b39d6b0173a6a1b6aab2881a607b6e8588ff124d4616f95fb944c
-  data.tar.gz: 20eedf19bb73922d482537bfe8411bf8b8816001a4df6d59130f08558d80e14d2bfd968c9029d176cae056b0a4d9b1c36d16fb4012329073cceb490846462f7a
+  metadata.gz: '0922a2a6d6061f7472ac55de8c6e0172a8ce8d2939f8f144f71493b148248fb597c07d916c67dc9a4549d35a6e65b7757e5cb3f0d0ae79a7a39f7b35306bbb22'
+  data.tar.gz: e5b5b898ca11ef0bdc077d8cfb78d6575bc7c9a6b712f2a0ccbd2a3f338704d66883c730e32651af1776746b751b996c723b29af80044d8cd44944f9b18df883

data/lib/canon/comparison/ruby_object_comparator.rb

@@ -24,6 +24,15 @@ module Canon
           return Comparison::UNEQUAL_TYPES
         end
 
+        # FAST PATH: deep-equal objects in identical key order — every
+        # nested comparison is trivially equivalent and no key-order
+        # difference exists (strict key_order compares insertion
+        # order, which keys == keys preserves). Most object pairs in
+        # similar documents hit this and skip the path-building walk.
+        if obj1 == obj2 && (!obj1.is_a?(Hash) || obj1.keys == obj2.keys)
+          return Comparison::EQUIVALENT
+        end
+
         case obj1
         when Hash
           compare_hashes(obj1, obj2, opts, differences, path)

data/lib/canon/formatters/json_formatter.rb

@@ -23,8 +23,12 @@ module Canon
       def self.sort_json_keys(obj)
         case obj
         when Hash
-          obj.transform_values { |v| sort_json_keys(v) }
-            .sort.to_h
+          # Single pass: sorted keys once, one rebuilt hash. The
+          # previous form (transform_values + sort.to_h) rebuilt every
+          # container twice per canonicalization.
+          sorted = {}
+          obj.keys.sort.each { |key| sorted[key] = sort_json_keys(obj[key]) }
+          sorted
         when Array
           obj.map { |item| sort_json_keys(item) }
         else

data/lib/canon/formatters/yaml_formatter.rb

@@ -23,8 +23,12 @@ module Canon
       def self.sort_yaml_keys(obj)
         case obj
         when Hash
-          obj.transform_values { |v| sort_yaml_keys(v) }
-            .sort.to_h
+          # Single pass: sorted keys once, one rebuilt hash (the
+          # transform_values + sort.to_h form rebuilt every container
+          # twice).
+          sorted = {}
+          obj.keys.sort.each { |key| sorted[key] = sort_yaml_keys(obj[key]) }
+          sorted
         when Array
           obj.map { |item| sort_yaml_keys(item) }
         else

data/lib/canon/html/data_model.rb

@@ -134,8 +134,7 @@ module Canon
             keep: Canon::Xml::WhitespacePolicy.keep_html_text?(
               node.content,
               parent_name: node.parent.is_a?(Nokogiri::XML::Element) ? node.parent.name : nil,
-              inline_significant:
-                Canon::Comparison::WhitespaceSensitivity.inline_whitespace_significant?(node),
+              text_node: node,
             ),
           )
         when Nokogiri::XML::Comment

data/lib/canon/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Canon
-  VERSION = "0.3.16"
+  VERSION = "0.3.18"
 end

data/lib/canon/xml/whitespace_policy.rb

@@ -61,13 +61,18 @@ element_parent: true)
       # so it is checked explicitly).
       HTML_WHITESPACE_SENSITIVE_TAGS = %w[pre code textarea script style].freeze
 
-      def keep_html_text?(content, parent_name:, inline_significant: false)
+      def keep_html_text?(content, parent_name:, text_node: nil)
         return true unless content.match?(STRIP_ONLY)
         return true if content.include?(" ")
 
         parent_name = parent_name.to_s.downcase
         return true if HTML_WHITESPACE_SENSITIVE_TAGS.include?(parent_name)
-        return true if inline_significant
+
+        # Computed last: the sibling scan is O(siblings), so it must
+        # not run for the content-bearing text nodes that fail the
+        # whitespace-only check above.
+        return true if text_node &&
+          Canon::Comparison::WhitespaceSensitivity.inline_whitespace_significant?(text_node)
 
         false
       end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: canon
 version: !ruby/object:Gem::Version
-  version: 0.3.16
+  version: 0.3.18
 platform: ruby
 authors:
 - Ribose Inc.