moxml 0.5.31 → 0.5.32

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 32f1034201de38bed35a6b1496d10f96a580f0e4c92517c2253ded0631331439
-  data.tar.gz: bb3d697a73fbac699e00b51905ef446f9837b9ee649785bb1ed4df2027870e78
+  metadata.gz: 627111c496efca4687eabeeb78989d611d3d308adc5ed32443c7058da51c57c9
+  data.tar.gz: 86d1cc6c471d07084e0d2e0de9412b286b5486b202f9fe677b7417f514c1fe55
 SHA512:
-  metadata.gz: 3a854ab201001d88494c37d79b163bbf4246880702f5c7d61cf52fd41128459be2a1e8c4601c894852b6aa73e33efca3c471ebfdbf9a126120afa698ab28aff1
-  data.tar.gz: 531a60b0d52e84f4f0df5a026d51b5eeed850b56da1c1eea23297c2c2fb331efa4047a9dc73ef21a430998035aad7a58fe8adb7905047f2001f5b941d07353fc
+  metadata.gz: f1108c98dc8af92913605dd08b84eb50bbfec92a5dd286e15656c3ecd57937ce56bc186c9393a740f6bd8a8316f06fb059db226f20ffc9f0b84cfb433e1a532d
+  data.tar.gz: 02d109a32659b30fc34ac5d6d3f01bc0d32d9e1ca6a38cfdc2fe3880a02c779ce5ff41fa2f0eda04adc2a9ccd6702f372d29b4d2a6fe33faeb0456cb8852d26b

data/docs/_pages/adapters/leptris.adoc

@@ -74,10 +74,17 @@ so script bodies with markup characters round-trip through a strict
 XML reparse. The adapter stands the entity-marker pipeline down for
 HTML parses — the engine decodes HTML entities directly into text.
 
-Measured on 1.9.105 (named-entity decoding fixed, engine #848):
+Measured on 1.9.105+ (named-entity decoding fixed, engine #848):
 parse 2.5-3.5x, serialize 4.1x, xpath 6-29x faster than Nokogiri —
 uniformly, including entity-laden pages.
 
+Since 1.9.118/1.9.121 the engine also handles WHATWG foreign
+content — embedded SVG and MathML in HTML parse into the tree
+(preserving `foreignObject` subtrees that Nokogiri drops), the
+source HTML doctype round-trips, and `template`/`colgroup` place
+correctly. Foreign-content camelCase names are currently
+lowercased (engine conformance note leptris/leptris#1013).
+
 Adapters without an engine HTML mode (Ox, Oga, REXML, LibXML) raise
 `Moxml::AdapterError`; Nokogiri parses through libxml2's HTML mode.
 

data/lib/moxml/adapter/leptris/document_parts.rb

@@ -64,6 +64,14 @@ module Moxml
             texts << child
             attachments.set(doc, :document_text, texts)
             child
+          when ::Leptris::XML::Comment
+            # The tree model supports document comments (they parse
+            # and serialize, libleptris 1.9.3 #578) but the engine
+            # has no add entry yet (leptris/leptris#1032).
+            raise Moxml::NotImplementedError.new(
+              "Adding document-level comments requires an engine entry (leptris/leptris#1032)",
+              feature: "add_document_child", adapter: "Leptris",
+            )
           else
             raise Moxml::DocumentStructureError.new(
               "Unsupported document child: #{child.class}",

data/lib/moxml/adapter/leptris/serialize.rb

@@ -33,7 +33,13 @@ module Moxml
           # Entity restoration belongs to the wrapper layer
           # (Node#to_xml runs adapter.restore_entities for every
           # adapter); doing it here scanned the output a second time.
-          xml = normalize_serialization(raw_serialize(node, options), options)
+          xml = if native_expand?(node, options)
+                  opts = options.dup
+                  opts[:__expand_handled_natively] = true
+                  normalize_serialization(raw_serialize(node, opts), opts)
+                else
+                  normalize_serialization(raw_serialize(node, options), options)
+                end
           # The binding's FFI strings come back binary-tagged; the
           # engine encoded the bytes per this option, so tag them.
           xml.force_encoding(options[:encoding]) if options[:encoding]
@@ -47,7 +53,32 @@ module Moxml
 
         # The binding's element face with all-default options — the
         # frozen splat keeps the hot argless path allocation-free.
+        # Native expand-empty (engine #882, libleptris 1.9.95,
+        # bindings 1.9.144): empty elements emit <a></a> through a
+        # C-side ext entry — the Ruby full-output regex rewrite and
+        # its "/>" probe scan drop out of the element path. The
+        # document face does not expose the option yet; documents
+        # keep the Ruby pass.
+        EXPAND_EMPTY_NATIVE =
+          Gem::Version.new(::Leptris::VERSION) >= Gem::Version.new("1.9.144")
+
         ELEMENT_DEFAULT_KWARGS = { indent: 0, no_decl: true, encoding: nil }.freeze
+        ELEMENT_EXPAND_KWARGS =
+          if EXPAND_EMPTY_NATIVE
+            { indent: 0, no_decl: true, encoding: nil, expand_empty: true }.freeze
+          else
+            ELEMENT_DEFAULT_KWARGS
+          end.freeze
+
+        # Elements (not documents) with expand_empty and NO
+        # indent-unit string: the C ext entry handles expansion. The
+        # element unit serializer does not take the flag — those keep
+        # the Ruby pass.
+        def native_expand?(node, options)
+          EXPAND_EMPTY_NATIVE && options[:expand_empty] &&
+            !node.is_a?(::Leptris::XML::Document) &&
+            !options[:indent_text].is_a?(String)
+        end
 
         def raw_serialize(node, options)
           # CDATA must precede Text in this chain: CDATA < Text in the
@@ -79,10 +110,12 @@ module Moxml
           # per call; the wrapper force-tags the string either way.
           indent = options.fetch(:indent, 0)
           encoding = options[:encoding] == "UTF-8" ? nil : options[:encoding]
+          native_expand = native_expand?(node, options)
           xml = if indent.zero? && encoding.nil?
-                  node.to_xml(**ELEMENT_DEFAULT_KWARGS)
+                  node.to_xml(**(native_expand ? ELEMENT_EXPAND_KWARGS : ELEMENT_DEFAULT_KWARGS))
                 else
                   kwargs = { indent: indent, no_decl: true, encoding: encoding }
+                  kwargs[:expand_empty] = true if native_expand
                   if INDENT_UNIT_SUPPORTED && options[:indent_text].is_a?(String)
                     kwargs[:indent_text] = options[:indent_text]
                   end
@@ -119,7 +152,9 @@ module Moxml
           # The libxml2-layout serializer (>= 1.9.42) keeps attribute
           # apostrophes literal; older engines escaped them.
           needs_apos = !LIBXML2_LAYOUT_PARITY && xml.include?("&apos;")
-          needs_expand = options[:expand_empty] && xml.include?("/>")
+          needs_expand = options[:expand_empty] &&
+            !options[:__expand_handled_natively] &&
+            xml.include?("/>")
           # Corruption guards: the 1-char ampersand probe is ~1µs
           # (memchr-class); the raw-< scan runs only on builds that
           # still carry the parse race (leptris-ruby#131).

data/lib/moxml/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Moxml
-  VERSION = "0.5.31"
+  VERSION = "0.5.32"
 end

data/spec/moxml/adapter/leptris_spec.rb

@@ -374,6 +374,37 @@ RSpec.describe Moxml::Adapter::Leptris do
       expect(reparsed.errors).to be_empty
     end
 
+    it "preserves foreign content (SVG/MathML) in HTML documents" do
+      doc = ctx.parse_html(%(<p>a</p><svg viewBox="0 0 1 1"><circle r="1"/></svg><math><mi>a</mi></math>))
+      body = doc.root.children.find(&:element?)
+      names = body.children.select(&:element?).map(&:name)
+      expect(names).to include("svg", "math")
+      expect(body.at_xpath(".//circle")["r"]).to eq("1")
+      expect(body.at_xpath(".//mi").text).to eq("a")
+    end
+
+    it "preserves foreignObject content Nokogiri drops (name lowercased)" do
+      # WHATWG keeps foreign-content camelCase (foreignObject,
+      # viewBox); the engine currently lowercases like it does HTML
+      # names, but preserves the subtree — Nokogiri drops it
+      # entirely. Pins current behavior; engine conformance note
+      # filed for the adjust-tables.
+      doc = ctx.parse_html(%(<svg><foreignObject><p>x</p></foreignObject></svg>))
+      out = doc.to_xml
+      expect(out).to include("<foreignobject>")
+      expect(out).to include("<p>x</p>")
+    end
+
+    it "round-trips an HTML doctype with external identifiers" do
+      doc = ctx.parse_html(%(<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN" "x.dtd"><p>a</p>))
+      expect(doc.to_xml).to include(%(<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN" "x.dtd">))
+    end
+
+    it "preserves template element placement" do
+      doc = ctx.parse_html(%(<div><template><p>t</p></template></div>))
+      expect(doc.at_xpath("//template/p")&.text).to eq("t")
+    end
+
     it "round-trips a parsed tree through mutation" do
       doc = ctx.parse_html(%(<ul><li>a<li>b</ul>))
       doc.at_xpath("//ul").add_child(doc.create_element("li"))

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: moxml
 version: !ruby/object:Gem::Version
-  version: 0.5.31
+  version: 0.5.32
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-09-12 00:00:00.000000000 Z
+date: 2026-09-13 00:00:00.000000000 Z
 dependencies: []
 description: |
   Moxml is a unified XML manipulation library that provides a common API