canon 0.3.63 → 0.3.64

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 0a25343915041867345d470570eb9a84d2266e4b95b89b730bee32a0a99b112f
-  data.tar.gz: 501f8c0e6c7430611a0c68c3b31b2d30ce63ec411bd6426e854fdb9c57a4f6ac
+  metadata.gz: b1124e218f58411b56e1cc00ab2e8047312e0aee560b3383bf1f7c394d6deb4f
+  data.tar.gz: ed5e13419a948b20edb312d242e21e8cba380d64db4ef76ca25595a6e07cb930
 SHA512:
-  metadata.gz: df2fb54642dffbc16670860f406bfce3e76d4665772d5d0394dd65762157ce821ab8d315c8a761abfd3d6ac9a72467039549fccaedfc5ea0ac7596852a898e28
-  data.tar.gz: 31781f1ef8b498980a1ca4036ebf9337f6716744f9f6d6e11519aa5ca4d0356d18c6679bc9091a10280b466abab0471e199c9db0d5c4138a27e992dff2af5623
+  metadata.gz: 1b49eb0267c0a8a42219d70686a722de07caa6576d46cec23f84c6c0f6c32d28eff247645747ff05f71d2ba7fb332672f36507a2715f3e732cf6cd99602d1aa9
+  data.tar.gz: 72ace1dd41a1fa94c1fdf51fbc6d807f752ab7930ddd7a5870724bf40ec24142d8d40b7a1d82f25e19e72ac373f7f06e4e7a26774821b90e1ccffc5d093bf62a

data/CLAUDE.md

@@ -143,6 +143,8 @@ Engine parity is complete through libleptris 1.9.8 / leptris-ruby 1.9.33 / moxml
 
 Engine A/B testing: `CANON_XML_BACKEND=nokogiri bundle exec rspec` (or `=moxml` to force leptris when it isn't the resolved default). The default suite must stay green under BOTH values; the only expected pendings are the upstream-tracked ones. The benchmark header (`rake performance:quick`) reports the active engine.
 
+Nokogiri is an OPTIONAL runtime dependency (dropped from the gemspec): the XML engine runs on moxml's leptris adapter, which ships native gems for every platform nokogiri does not (aarch64-mingw-ucrt etc.). `Canon::NokogiriLoader` lazy-loads nokogiri only for HTML support and the raw `:nokogiri` engine, turning absence into an actionable `Canon::Error` at use time. The contract is gated by `CANON_SIMULATE_NO_NOKOGIRI=1 bundle exec rspec` (a load-path blocker makes every `require "nokogiri"` fail like a missing gem; nokogiri-only spec trees are excluded via `.exclude_pattern` + `:requires_nokogiri` tags in spec_helper). Specs constructing raw Nokogiri objects guard their require with `if Canon::NokogiriLoader.available?`. Adding a new raw `Nokogiri::` constant reference in lib requires either a `defined?(Nokogiri)` guard or a lazy `NokogiriLoader.require!` — the nokogiri-less leg exists to enforce this.
+
 ### C14N Engines
 
 `Canon::Xml::C14n.canonicalize` uses leptris' native C-side C14N 1.1 by DEFAULT (since libleptris 1.9.178 / gem 1.9.178.0) — 23x faster than the Ruby processor through canon's API (1MB document). The native parse runs `noblanks: true` (compact bytes, matching the Ruby lane's product), relative namespace URIs raise identically, and document-level PIs serialize in spec-correct document order (the Ruby processor's root-first order was non-conformant for prolog PIs). `CANON_C14N_BACKEND=ruby` forces the Ruby processor. `canonicalize_subset` and `with_comments: true` always use the Ruby processor. The parity spec (`spec/canon/xml/c14n_engine_parity_spec.rb`) pins byte-identity; its pending cases track intentional divergences (document-order PIs) and the canon-side CR-reference finding.

data/lib/canon/comparison/html_comparator.rb

@@ -1,7 +1,7 @@
 # frozen_string_literal: true
 
 require "set"
-require "nokogiri" unless RUBY_ENGINE == "opal"
+Canon::NokogiriLoader.require!("HTML comparison")
 
 module Canon
   module Comparison

data/lib/canon/comparison/html_parser.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require "nokogiri" unless RUBY_ENGINE == "opal"
+Canon::NokogiriLoader.require!("HTML comparison")
 
 module Canon
   module Comparison

data/lib/canon/comparison/pipeline.rb

@@ -252,8 +252,12 @@ module Canon
           case obj
           when String
             obj
-          when Nokogiri::XML::Document, Nokogiri::HTML::Document,
-               Nokogiri::XML::DocumentFragment, Nokogiri::HTML::DocumentFragment
+          when *(if defined?(Nokogiri)
+                   [Nokogiri::XML::Document, Nokogiri::HTML::Document,
+                    Nokogiri::XML::DocumentFragment, Nokogiri::HTML::DocumentFragment]
+                 else
+                   []
+                 end)
             obj.to_html
           else
             if Canon::XmlParsing.xml_node?(obj) || obj.is_a?(Canon::Xml::Node)

data/lib/canon/comparison.rb

@@ -1,7 +1,6 @@
 # frozen_string_literal: true
 
 require "moxml"
-require "nokogiri" unless RUBY_ENGINE == "opal"
 
 module Canon
   # Comparison module for XML, HTML, JSON, and YAML documents

data/lib/canon/diff_formatter/diff_detail_formatter/dimension_formatter.rb

@@ -424,7 +424,7 @@ expand_difference: false)
           case node
           when Canon::Xml::Node
             node.node_type == :element
-          when Nokogiri::XML::Node
+          when *(defined?(Nokogiri) ? [Nokogiri::XML::Node] : [])
             node.element?
           else
             false

data/lib/canon/diff_formatter/diff_detail_formatter/location_extractor.rb

@@ -64,7 +64,7 @@ module Canon
 
           while current
             name = case current
-                   when Canon::Xml::Node, Nokogiri::XML::Node
+                   when Canon::Xml::Node, *(defined?(Nokogiri) ? [Nokogiri::XML::Node] : [])
                      current.name
                    else
                      break
@@ -75,13 +75,13 @@ module Canon
             parts.unshift("#{name}[#{index}]")
 
             current = case current
-                      when Canon::Xml::Node, Nokogiri::XML::Node
+                      when Canon::Xml::Node, *(defined?(Nokogiri) ? [Nokogiri::XML::Node] : [])
                         current.parent
                       else
                         break
                       end
 
-            break if current.is_a?(Nokogiri::XML::Document) ||
+            break if (defined?(Nokogiri) && current.is_a?(Nokogiri::XML::Document)) ||
               current.is_a?(Canon::Xml::Nodes::RootNode)
           end
 
@@ -95,17 +95,17 @@ module Canon
         # @return [Integer] 1-based index
         def self.calculate_sibling_index(node, name)
           parent = case node
-                   when Canon::Xml::Node, Nokogiri::XML::Node
+                   when Canon::Xml::Node, *(defined?(Nokogiri) ? [Nokogiri::XML::Node] : [])
                      node.parent
                    end
 
           return 1 unless parent
 
           siblings = case parent
-                     when Canon::Xml::Node, Nokogiri::XML::Node
+                     when Canon::Xml::Node, *(defined?(Nokogiri) ? [Nokogiri::XML::Node] : [])
                        parent.children.select do |n|
                          case n
-                         when Canon::Xml::Node, Nokogiri::XML::Node
+                         when Canon::Xml::Node, *(defined?(Nokogiri) ? [Nokogiri::XML::Node] : [])
                            n.name == name
                          else
                            false

data/lib/canon/diff_formatter.rb

@@ -590,7 +590,7 @@ module Canon
           ),
         ]
       when :html
-        require "nokogiri"
+        Canon::NokogiriLoader.require!("HTML diff formatting")
         [
           parse_and_format_html(expected),
           parse_and_format_html(actual),
@@ -623,10 +623,11 @@ module Canon
     # @param html [Object] HTML content
     # @return [String] Formatted HTML
     def parse_and_format_html(html)
-      return html.to_html if html.is_a?(Nokogiri::HTML::Document) ||
+      return html.to_html if (defined?(Nokogiri) &&
+        html.is_a?(Nokogiri::HTML::Document)) ||
         html.is_a?(Nokogiri::HTML5::Document)
 
-      require "nokogiri"
+      Canon::NokogiriLoader.require!("HTML normalization")
       Nokogiri::HTML(html).to_html
     rescue StandardError
       html.to_s
@@ -1080,7 +1081,7 @@ show_received: true)
     # Normalizes attribute order, void elements, and optional end tags consistently.
     # Falls back to the original string on any parse error.
     def safe_html_normalize(doc)
-      require "nokogiri"
+      Canon::NokogiriLoader.require!("HTML normalization")
       Nokogiri::HTML5(doc.to_s).to_html(encoding: "UTF-8")
     rescue StandardError
       doc.to_s

data/lib/canon/formatters/html_formatter.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require "nokogiri" unless RUBY_ENGINE == "opal"
+Canon::NokogiriLoader.require!("HTML formatting")
 
 module Canon
   module Formatters

data/lib/canon/formatters/html_formatter_base.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require "nokogiri" unless RUBY_ENGINE == "opal"
+Canon::NokogiriLoader.require!("HTML formatting")
 require "set"
 
 module Canon

data/lib/canon/formatters/xml_formatter.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-require "nokogiri" unless RUBY_ENGINE == "opal"
-
 module Canon
   module Formatters
     class XmlFormatter

data/lib/canon/html/data_model.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require "nokogiri" unless RUBY_ENGINE == "opal"
+Canon::NokogiriLoader.require!("HTML support")
 
 module Canon
   module Html

data/lib/canon/nokogiri_loader.rb

@@ -0,0 +1,43 @@
+# frozen_string_literal: true
+
+module Canon
+  # Nokogiri is an OPTIONAL dependency.
+  #
+  # canon's XML engine runs on moxml's resolved adapter — leptris on every
+  # platform (leptris ships native gems for all of them, including platforms
+  # where nokogiri has no prebuilt). Nokogiri is required only by the raw
+  # `CANON_XML_BACKEND=nokogiri` engine and by HTML support (moxml has no
+  # HTML adapter). Everything here turns "nokogiri missing" into a loud,
+  # actionable Canon::Error at the moment a nokogiri-only feature is used,
+  # instead of a LoadError at gem-install or require time.
+  module NokogiriLoader
+    MSG = "requires the nokogiri gem, which is not installed. " \
+          "XML canonicalization, comparison, diffing, and pretty-printing " \
+          "work without it (moxml + leptris engine); install nokogiri only " \
+          "if you need HTML support or the CANON_XML_BACKEND=nokogiri engine."
+    private_constant :MSG
+
+    class << self
+      # Load nokogiri or raise a helpful Canon::Error naming the feature.
+      # No-op under Opal (matches the historical `unless RUBY_ENGINE`
+      # guards).
+      def require!(feature)
+        return if RUBY_ENGINE == "opal"
+
+        require "nokogiri"
+      rescue LoadError
+        raise Canon::Error, "#{feature} #{MSG}"
+      end
+
+      # True when nokogiri can be loaded (never raises).
+      def available?
+        return false if RUBY_ENGINE == "opal"
+
+        require "nokogiri"
+        true
+      rescue LoadError
+        false
+      end
+    end
+  end
+end

data/lib/canon/pretty_printer/html.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require "nokogiri" unless RUBY_ENGINE == "opal"
+Canon::NokogiriLoader.require!("HTML pretty-printing")
 require "stringio"
 
 module Canon

data/lib/canon/pretty_printer/xml.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-require "nokogiri" unless RUBY_ENGINE == "opal"
-
 module Canon
   module PrettyPrinter
     class Xml
@@ -28,6 +26,7 @@ module Canon
       private
 
       def nokogiri_format(xml_string)
+        Canon::NokogiriLoader.require!("Nokogiri-engine XML pretty-printing")
         doc = Nokogiri::XML(xml_string, &:noblanks)
         if @indent_type == "tab"
           doc.to_xml(indent: 1, indent_text: "\t", encoding: "UTF-8")

data/lib/canon/pretty_printer/xml_normalized.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-require "nokogiri" unless RUBY_ENGINE == "opal"
-
 module Canon
   module PrettyPrinter
     # Mixed-content-aware XML serializer for diff display preprocessing.
@@ -162,8 +160,10 @@ module Canon
               end
         lines = []
 
-        if !@html_mode && doc.version
-          enc = doc.encoding ? " encoding=\"#{doc.encoding}\"" : ""
+        # Moxml documents (the nokogiri-less fallback) answer neither
+        # #version nor #encoding.
+        if !@html_mode && doc.respond_to?(:version) && doc.version
+          enc = doc.respond_to?(:encoding) && doc.encoding ? " encoding=\"#{doc.encoding}\"" : ""
           lines << "<?xml version=\"#{doc.version}\"#{enc}?>"
         end
 

data/lib/canon/rspec_matchers.rb

@@ -5,8 +5,11 @@ require "canon/comparison"
 require "canon/diff_formatter"
 require "canon/config"
 require "canon/rebaseliner"
-require "canon/pretty_printer/xml"
-require "canon/pretty_printer/html"
+# PrettyPrinter::Xml/Html are NOT required here: Html loads nokogiri at
+# its top (lazy via the Canon::PrettyPrinter autoload registry), and a
+# nokogiri-less host must be able to load canon's matchers — the HTML
+# printer surfaces the actionable Canon::Error only when the HTML matcher
+# actually formats output.
 
 begin
   require "rspec/expectations"

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

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require "nokogiri" unless RUBY_ENGINE == "opal"
+Canon::NokogiriLoader.require!("HTML tree diffing")
 
 module Canon
   module TreeDiff

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

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-require "nokogiri" unless RUBY_ENGINE == "opal"
-
 module Canon
   module TreeDiff
     module Adapters
@@ -49,7 +47,11 @@ module Canon
             return to_tree_from_canon_comment(node)
           end
 
-          # Fallback to Nokogiri (legacy support)
+          # Fallback to raw engine documents (legacy support). Nokogiri
+          # is optional — under the moxml/leptris engine a raw Nokogiri
+          # object cannot exist, so the constants stay unresolved.
+          return unless defined?(Nokogiri)
+
           case node
           when Nokogiri::XML::Document
             # Start from root element
@@ -70,6 +72,7 @@ module Canon
         # @param doc [Nokogiri::XML::Document] Optional document to use
         # @return [Nokogiri::XML::Document, Nokogiri::XML::Element]
         def from_tree(tree_node, doc = nil)
+          Canon::NokogiriLoader.require!("XML tree reconstruction")
           doc ||= Nokogiri::XML::Document.new
 
           element = build_element(tree_node, doc)

data/lib/canon/validators/base_validator.rb

@@ -27,7 +27,7 @@ module Canon
         line = nil
         column = nil
 
-        if error.is_a?(Nokogiri::XML::SyntaxError)
+        if defined?(Nokogiri) && error.is_a?(Nokogiri::XML::SyntaxError)
           line = error.line
           column = error.column
         elsif error.message =~ /line[:\s]+(\d+)/i

data/lib/canon/validators/html_validator.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require "nokogiri" unless RUBY_ENGINE == "opal"
+Canon::NokogiriLoader.require!("HTML validation")
 
 module Canon
   module Validators

data/lib/canon/validators/xml_validator.rb

@@ -1,14 +1,13 @@
 # frozen_string_literal: true
 
-require "nokogiri" unless RUBY_ENGINE == "opal"
-
 module Canon
   module Validators
     # Validator for XML input
     #
-    # Validates XML input using Nokogiri's strict parsing mode.
-    # Raises detailed ValidationError with line and column information
-    # when malformed XML is detected.
+    # Validates XML input with strict parsing when nokogiri is available
+    # (detailed line/column errors). Without nokogiri, validation parses
+    # through the active engine (moxml/leptris), which still rejects
+    # malformed input — only the location detail is coarser.
     class XmlValidator < BaseValidator
       # Validate XML input
       #
@@ -18,7 +17,19 @@ module Canon
       def self.validate!(input)
         return if input.nil? || input.strip.empty?
 
-        # Parse with strict error handling
+        if Canon::NokogiriLoader.available?
+          nokogiri_validate!(input)
+        else
+          engine_validate!(input)
+        end
+      end
+
+      # Parse with strict error handling
+      #
+      # @param input [String] The XML string to validate
+      # @raise [Canon::ValidationError] If XML is malformed
+      # @return [void]
+      def self.nokogiri_validate!(input)
         Nokogiri::XML(input) do |config|
           config.strict.nonet
         end
@@ -33,6 +44,25 @@ module Canon
         )
       end
 
+      # Validate by parsing through the engine selected by
+      # Canon::XmlBackend (never raw nokogiri).
+      #
+      # @param input [String] The XML string to validate
+      # @raise [Canon::ValidationError] If XML is malformed
+      # @return [void]
+      def self.engine_validate!(input)
+        Canon::XmlParsing.parse(input)
+      rescue StandardError => e
+        location = extract_location(e)
+        raise Canon::ValidationError.new(
+          e.message.split("\n").first,
+          format: :xml,
+          line: location[:line],
+          column: location[:column],
+          details: extract_details(e),
+        )
+      end
+
       # Extract additional error details
       #
       # @param error [Nokogiri::XML::SyntaxError] The syntax error

data/lib/canon/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Canon
-  VERSION = "0.3.63"
+  VERSION = "0.3.64"
 end

data/lib/canon/xml/data_model.rb

@@ -1,6 +1,5 @@
 # frozen_string_literal: true
 
-require "nokogiri" unless RUBY_ENGINE == "opal"
 require "set"
 
 module Canon
@@ -119,6 +118,7 @@ module Canon
       # --- Nokogiri path ---
 
       def self.from_nokogiri_xml(xml_string, preserve_whitespace:)
+        Canon::NokogiriLoader.require!("Nokogiri-engine XML parsing")
         doc = Nokogiri::XML(xml_string, &:nonet)
         check_for_relative_namespace_uris(doc)
         result = build_from_nokogiri(doc,

data/lib/canon/xml/sax/nokogiri_driver.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require "nokogiri" unless RUBY_ENGINE == "opal"
+Canon::NokogiriLoader.require!("the Nokogiri SAX driver")
 
 module Canon
   module Xml

data/lib/canon/xml_backend.rb

@@ -54,6 +54,12 @@ module Canon
                 "Must be one of: #{VALID_BACKENDS.join(', ')}"
         end
 
+        # The raw :nokogiri engine executes raw Nokogiri::XML calls —
+        # with nokogiri optional, fail here (at selection) with the
+        # actionable message rather than NameError at first parse.
+        Canon::NokogiriLoader.require!("CANON_XML_BACKEND=nokogiri") if
+          backend == :nokogiri
+
         backend
       end
 

data/lib/canon.rb

@@ -2,7 +2,11 @@
 
 require "canon/version"
 require "canon/errors"
-require "nokogiri" unless RUBY_ENGINE == "opal"
+# Nokogiri is optional: loaded lazily by Canon::NokogiriLoader only for
+# HTML support and the raw CANON_XML_BACKEND=nokogiri engine. XML runs on
+# moxml's resolved adapter (leptris), which ships native gems for every
+# platform nokogiri does not.
+require "canon/nokogiri_loader"
 require "canon/xml_backend"
 require "canon/xml_parsing"
 require "canon/yaml_backend"

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: canon
 version: !ruby/object:Gem::Version
-  version: 0.3.63
+  version: 0.3.64
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-09-20 00:00:00.000000000 Z
+date: 2026-09-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: diff-lcs
@@ -44,42 +44,28 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.9.193
+        version: 1.9.216
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 1.9.193
+        version: 1.9.216
 - !ruby/object:Gem::Dependency
   name: moxml
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.5.60
+        version: 0.5.71
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.5.60
-- !ruby/object:Gem::Dependency
-  name: nokogiri
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0'
+        version: 0.5.71
 - !ruby/object:Gem::Dependency
   name: paint
   requirement: !ruby/object:Gem::Requirement
@@ -156,14 +142,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.6.5
+        version: 0.6.16
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.6.5
+        version: 0.6.16
 description: |-
   Canon provides canonicalization and pretty-printing for various serialization
   formats (XML, HTML, JSON, YAML), producing standardized forms suitable for
@@ -362,6 +348,7 @@ files:
 - lib/canon/html/data_model.rb
 - lib/canon/html/nokogiri_support.rb
 - lib/canon/json_parsing.rb
+- lib/canon/nokogiri_loader.rb
 - lib/canon/options.rb
 - lib/canon/options/cli_generator.rb
 - lib/canon/options/registry.rb