moxml 0.5.16 → 0.5.18

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 39cfbe08ea2a0e92f2ba89e83beacf564401765e964807dc4d73ca23d867af38
-  data.tar.gz: cdd5d499fd4b3233a6c4d2774b188aef2339f341f7c7a7112a82ac392ec4e53b
+  metadata.gz: c793d84315c4005a5141698aeb716c001d23b7157b768c77cbdb3539a4b08f53
+  data.tar.gz: 35c1d60e6d3a8e17e9768c733d028962bcb6e5798e3c8528b18deaebd855ae5f
 SHA512:
-  metadata.gz: 894d6bacdd4f26dbb2a59b0684b29e8ca68b39b66f841a5cd11005351eff38404e745b47d58bb860c898d3da92e1e2a769210938f26cac2f88cac34cc887de76
-  data.tar.gz: c5b7b7c9de4f2362e910b740ce4cf6185a72773f39f31a3dfc17fbd7436f65f542e66ff099900bf9073e87bd5975568cb3833ac658115ea5ee21a4218719ce90
+  metadata.gz: '092b193420e9117adbd6dade8718474c9a72de401eb0b7a741d36639bbd7e9285c7e53d55f96b482b17bea3382d1da3adb1cae6bbc7b787c552859d80377db39'
+  data.tar.gz: 18a342a60a7dda872730dd2a9158aa852948596e21fc4613961d0336ece2f12977d02fe434e7fa64429f3ce73015a4434285f3f052425c4eb6d88c20a9571db7

data/docs/_pages/adapter-protocol.adoc

@@ -18,7 +18,9 @@ code disagree.
 == Method groups
 
 Parse and document lifecycle::
-`parse(xml, options, context)`, `create_document`, `create` factories
+`parse(xml, options, context)` — `noblanks: true` drops
+whitespace-only text nodes where the engine has the flag
+(Nokogiri, leptris), `create_document`, `create` factories
 (`create_native_element`, `create_native_text`, `create_native_cdata`,
 `create_native_comment`, `create_native_processing_instruction`,
 `create_native_declaration`, `create_native_doctype`,
@@ -55,7 +57,10 @@ the tracked native; wrappers re-key through `refresh_native!`.
 
 Serialization::
 `serialize(node, options)`. Options: `indent`, `encoding`,
-`declaration`, `no_declaration`, `expand_empty`, `line_ending`.
+`declaration`, `no_declaration`, `expand_empty`, `line_ending`,
+`indent_text` (the indent-unit string — Nokogiri and
+leptris >= 1.9.45; leptris text-bearing leaves wait on
+leptris/leptris#658).
 Entity-marker restoration is owned by the wrapper layer
 (`Node#to_xml`), guarded by `entity_bearing?`.
 

data/docs/_pages/adapters/leptris.adoc

@@ -30,6 +30,11 @@ ctx.config.adapter_name        # => :leptris / :nokogiri / :rexml (Opal)
 
 == Parse options
 
+`noblanks: true`::
+Drop whitespace-only text nodes (libxml2 `XML_PARSE_NOBLANKS`
+parity, issue #153) — the parse shape canon's pretty-printer
+compares against `Nokogiri::XML(xml, &:noblanks)`.
+
 `readonly: true`::
 The binding memoizes every read and refuses mutations — the
 parse-and-read lifecycle for multi-pass consumers (comparison, diff,
@@ -62,16 +67,18 @@ implementation over one `leptris_node_traverse` call.
 
 == Known limitations (tracked upstream)
 
-* Tab indentation (`indent_text`) needs a serialize-time parameter in
-  libleptris; post-hoc rewriting would corrupt mixed content.
-* Child PIs serialize inline (engine emits `<r>\n<?pi x?>  <e/>`
-  rather than libxml2's one-PI-per-indented-line layout); safe
-  post-hoc rewriting is impossible in mixed content.
-* DOCTYPE internal subsets serialize inline; libxml2 separates the
-  declarations with newlines.
-* Document-level PIs are read-only natives (`target=`/`data=`/
-  `unlink` rejected); write-through and removal are pending C
-  surface (leptris/leptris#612).
+* Tab indentation honors the indent-unit string on leptris >= 1.9.45
+  (`indent_text:`), byte-identical to Nokogiri on element-only
+  trees; text-bearing leaves still emit spaces until the engine fix
+  (leptris/leptris#658).
 * A PI preceding the DOCTYPE reorders on serialization — libleptris
   normalizes DOCTYPE-first and the document node does not record its
   relative position.
+
+Resolved with leptris-ruby 1.9.42 (leptris/leptris#636/#612): child
+PIs serialize on their own indented lines, DOCTYPE internal subsets
+newline-separate their declarations (moxml re-formats the subset to
+the libxml2 layout), the stray trailing newline after non-ASCII
+ending elements is gone engine-side, and document-level PIs accept
+`target=`/`content=`/removal with full round-trip. The moxml-side
+compat guards remain as no-ops for older floor bindings.

data/docs/_pages/conversion-apis.adoc

@@ -100,4 +100,9 @@ context.parse(src).to_xml(
 ) == Nokogiri::XML(src).to_xml(indent: 2, encoding: "UTF-8")   # => true
 ----
 
-Known residual divergences (engine-level, tracked upstream): tab indentation needs a serialize-time parameter in libleptris, and a PI preceding the DOCTYPE reorders (the engine normalizes DOCTYPE-first).
+Known residual divergences (engine-level, tracked upstream): a PI
+preceding the DOCTYPE reorders (the engine normalizes DOCTYPE-first),
+and tab indentation waits on the binding exposing an indent-unit
+parameter (leptris-ruby#109) — child-PI lines, DOCTYPE subset
+layout, and the non-ASCII trailing-newline quirk were resolved with
+leptris-ruby 1.9.42.

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

@@ -113,9 +113,75 @@ module Moxml
           return nil unless dt
 
           subset = dt.internal_subset if dt.class.method_defined?(:internal_subset)
+          subset = format_internal_subset(subset) if LIBXML2_LAYOUT_PARITY
           XmlEmitter.doctype_xml(dt.root_name, dt.public_id, dt.system_id, subset)
         end
 
+        # libxml2's DTD dump layout (leptris/leptris#636): newline
+        # after "[", one after every markup declaration, none after
+        # comments (they glue to both neighbors); an empty subset
+        # drops the brackets. Returns the INNER text for
+        # XmlEmitter.doctype_xml, nil when there is nothing to emit.
+        # The engine reports internal_subset as raw source text, so
+        # the declarations are re-tokenized — quote-aware, since an
+        # attribute default can contain ">".
+        def format_internal_subset(subset)
+          return nil if subset.nil? || subset.empty?
+
+          out = +"\n"
+          pos = 0
+          length = subset.length
+          while pos < length
+            start = subset.index("<", pos)
+            break if start.nil?
+
+            terminator, skip = if subset[start, 4] == "<!--"
+                                 ["-->", 4]
+                               elsif subset[start, 2] == "<?"
+                                 ["?>", 2]
+                               else
+                                 [nil, 0]
+                               end
+            if terminator
+              stop = subset.index(terminator, start + skip)
+              break if stop.nil?
+
+              item_end = stop + terminator.length
+            else
+              item_end = markup_decl_end(subset, start)
+              break if item_end.nil?
+            end
+            out << subset[start...item_end]
+            # Comments carry no trailing newline; declarations do.
+            out << "\n" unless subset[start, 4] == "<!--"
+            pos = item_end
+          end
+          out == "\n" ? nil : out
+        end
+
+        # End index of a markup declaration starting at `start`:
+        # the first ">" outside quotes.
+        QUOTE_CHARS = ['"', "'"].freeze
+        private_constant :QUOTE_CHARS
+
+        def markup_decl_end(subset, start)
+          quote = nil
+          i = start
+          length = subset.length
+          while i < length
+            ch = subset[i]
+            if quote
+              quote = nil if ch == quote
+            elsif QUOTE_CHARS.include?(ch)
+              quote = ch
+            elsif ch == ">"
+              return i + 1
+            end
+            i += 1
+          end
+          nil
+        end
+
         def default_declaration_xml(doc, options)
           encoding = options[:encoding] || doc.encoding
           encoding = "UTF-8" if encoding.to_s.empty?

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

@@ -63,14 +63,21 @@ module Moxml
           include_decl = options.fetch(:declaration) do
             options[:no_declaration] ? false : document_has_declaration?(node)
           end
-          xml = node.to_xml(
+          kwargs = {
             indent: options.fetch(:indent, 0),
             no_decl: !include_decl,
             encoding: options[:encoding],
-          )
+          }
+          if INDENT_UNIT_SUPPORTED && options[:indent_text].is_a?(String)
+            kwargs[:indent_text] = options[:indent_text]
+          end
+          xml = node.to_xml(**kwargs)
           # Element output always ends with the close tag — but the
           # engine's serializer appends a stray trailing newline when
-          # the element's last text child is non-ASCII.
+          # the element's last text child is non-ASCII (fixed engine
+          # side in 1.9.42; kept for older floor bindings). Note: the
+          # binding's indent_text kwarg is the display-form boolean,
+          # not Nokogiri's indent-unit string — do not forward it.
           xml.sub(/\n+\z/, "")
         end
 

data/lib/moxml/adapter/leptris.rb

@@ -31,6 +31,24 @@ module Moxml
       PREFIXED_ATTR_PREDICATES_NATIVE =
         Gem::Version.new(::Leptris::VERSION) >= Gem::Version.new("1.9.40")
 
+      # leptris/leptris#636 (leptris-ruby 1.9.42): libxml2-compatible
+      # pretty-print — child-PI lines, DOCTYPE internal-subset layout,
+      # no stray trailing newline after non-ASCII text, and the indent
+      # unit (`to_xml`'s indent_text takes the unit string). Older
+      # floor bindings treat indent_text as a display-form boolean, so
+      # the passthrough and the parity pins gate on this.
+      LIBXML2_LAYOUT_PARITY =
+        Gem::Version.new(::Leptris::VERSION) >= Gem::Version.new("1.9.42")
+
+      # leptris-ruby#109 (1.9.45): Element#to_xml takes the
+      # indent-unit string, so moxml's per-child document composition
+      # can carry `indent_text:` through. Text-bearing leaves still
+      # emit spaces until the engine fix (leptris/leptris#658); the
+      # boolean display form stays document-level and is never
+      # forwarded (the element face raises on it).
+      INDENT_UNIT_SUPPORTED =
+        Gem::Version.new(::Leptris::VERSION) >= Gem::Version.new("1.9.45")
+
       NO_PARSE_ERRORS = [].freeze
       private_constant :NO_PARSE_ERRORS
 
@@ -72,7 +90,7 @@ module Moxml
             ::Leptris::XML::Document.parse(
               processed,
               readonly: options[:readonly] == true,
-              options: dtdattr_parse_options(options),
+              options: parse_flags(options),
             )
           rescue ::Leptris::XML::ParseError => e
             # libleptris has no recovery mode that survives unclosed
@@ -95,11 +113,16 @@ module Moxml
           doc
         end
 
-        # nil unless DTDATTR is requested — the binding treats a nil
-        # options hash as plain defaults (defaults exclude ATTLIST
-        # defaults, matching libxml2/Nokogiri semantics).
-        def dtdattr_parse_options(options)
-          options[:dtdattr] == true ? ::Leptris::XML::ParseOptions.dtdattr : nil
+        # nil when no parse flag is requested — the binding treats a
+        # nil options hash as plain defaults (matching libxml2/
+        # Nokogiri semantics: blanks kept, no ATTLIST defaults).
+        # noblanks (issue #153) drops whitespace-only text nodes,
+        # libxml2 XML_PARSE_NOBLANKS parity.
+        def parse_flags(options)
+          flags = 0
+          flags |= ::Leptris::XML::ParseOptions::DTDATTR if options[:dtdattr] == true
+          flags |= ::Leptris::XML::ParseOptions::NOBLANKS if options[:noblanks] == true
+          flags.zero? ? nil : ::Leptris::XML::ParseOptions.new(flags)
         end
 
         def parse_errors(native_doc)

data/lib/moxml/adapter/nokogiri.rb

@@ -26,11 +26,13 @@ module Moxml
               ::Nokogiri::XML::DocumentFragment.parse(processed_xml) do |config|
                 config.strict.nonet
                 config.recover unless options[:strict]
+                config.noblanks if options[:noblanks]
               end
             else
               ::Nokogiri::XML(processed_xml, nil, "UTF-8") do |config|
                 config.strict.nonet
                 config.recover unless options[:strict]
+                config.noblanks if options[:noblanks]
               end
             end
           rescue ::Nokogiri::XML::SyntaxError => e
@@ -429,11 +431,13 @@ module Moxml
             custom_decl = "<?xml #{attrs.join(' ')}?>"
           end
 
-          result = node.to_xml(
+          serialize_kwargs = {
             indent: options[:indent],
             encoding: options[:encoding],
             save_with: save_options,
-          )
+          }
+          serialize_kwargs[:indent_text] = options[:indent_text] if options[:indent_text]
+          result = node.to_xml(**serialize_kwargs)
 
           if custom_decl
             result = "#{custom_decl}\n#{result}"

data/lib/moxml/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Moxml
-  VERSION = "0.5.16"
+  VERSION = "0.5.18"
 end

data/spec/moxml/adapter/leptris_spec.rb

@@ -157,6 +157,15 @@ RSpec.describe Moxml::Adapter::Leptris do
         selfclosing: %(<r><a/><b/><c>t</c><d/></r>),
       }
 
+      # leptris/leptris#636 (binding 1.9.42): child-PI lines and
+      # DOCTYPE internal-subset layout match libxml2.
+      if defined?(described_class::LIBXML2_LAYOUT_PARITY) && described_class::LIBXML2_LAYOUT_PARITY
+        cases[:pi_child] = %(<r><?pi data?><e/></r>)
+        cases[:pi_child_mixed] = %(<r>t<?pi d?><e>x</e>u</r>)
+        cases[:doctype_subset] = %(<?xml version="1.0"?><!DOCTYPE r [<!ELEMENT r (#PCDATA)>]><r>t</r>)
+        cases[:doctype_subset_multi] = %(<?xml version="1.0"?><!DOCTYPE r [<!ELEMENT r (#PCDATA)><!ATTLIST e a CDATA "d">]><r><e/></r>)
+      end
+
       cases.each do |name, source|
         target = Nokogiri::XML(source).to_xml(indent: 2, encoding: "UTF-8")
         output = ctx.parse(source).to_xml(
@@ -166,6 +175,43 @@ RSpec.describe Moxml::Adapter::Leptris do
       end
     end
 
+    it "round-trips document-level PI mutations through serialization" do
+      # leptris/leptris#612: parse-created document PIs carry doc
+      # linkage — the setters work and the tree round-trips.
+      doc = ctx.parse("<?pi-prolog before?><root/>")
+      pi = doc.children.to_a.first
+      pi.target = "renamed"
+      pi.content = "changed"
+      expect(doc.to_xml).to include("<?renamed changed?>")
+    end
+
+    it "matches raw Nokogiri with tab units on element-only trees" do
+      skip "requires the indent unit (leptris 1.9.45+)" unless described_class::INDENT_UNIT_SUPPORTED
+
+      source = %(<r><a><b/></a><c><d/><e/></c></r>)
+      target = Nokogiri::XML(source).to_xml(indent: 2, indent_text: "\t", encoding: "UTF-8")
+      output = ctx.parse(source).to_xml(
+        indent: 2, declaration: true, expand_empty: false,
+        encoding: "UTF-8", indent_text: "\t"
+      )
+      expect(output).to eq(target)
+    end
+
+    it "drops space-only text nodes with noblanks and matches Nokogiri (issue #153)" do
+      source = %(<a><t>1</t>    <n/></a>)
+      recipe = {
+        indent: 2, declaration: true, expand_empty: false, encoding: "UTF-8"
+      }
+      target = Nokogiri::XML(source, &:noblanks).to_xml(indent: 2, encoding: "UTF-8")
+
+      doc = ctx.parse(source, noblanks: true)
+      expect(doc.root.children.to_a.select(&:text?)).to be_empty
+      expect(doc.to_xml(recipe)).to eq(target)
+
+      via_nokogiri = Moxml.new(:nokogiri).parse(source, noblanks: true).to_xml(recipe)
+      expect(via_nokogiri).to eq(target)
+    end
+
     it "reports the tracked native for a re-added document PI" do
       doc = ctx.parse("<?pi-src orig?><root/>")
       moved = doc.children.to_a[0]

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: moxml
 version: !ruby/object:Gem::Version
-  version: 0.5.16
+  version: 0.5.18
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-08-30 00:00:00.000000000 Z
+date: 2026-08-31 00:00:00.000000000 Z
 dependencies: []
 description: |
   Moxml is a unified XML manipulation library that provides a common API