moxml 0.5.83 → 0.5.84

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 47eeb00aed4df2a6ea9474713ed65ba577889375e2ea7a49a8b36ee7fde86090
-  data.tar.gz: 947e4a25bd867ce42a52015ceea6fe576ded8b1ccf9a232cc5b1a69e1c22b166
+  metadata.gz: ff836e409dd2cf77368c7b26015e57de1b0af5e0678489e6e10452a6c70cc61a
+  data.tar.gz: a2a09b561505f90fc624089e925c3b41d4ab867668b00fd35f1132787e9dbfde
 SHA512:
-  metadata.gz: daaf3e2d077a2fee6b710d61b512a1e8c89922727cbf147aca2fb69147602aa3c50369bdc990cb385e3522296151102fa5498f0df5e11955d1b16424fa18c2b1
-  data.tar.gz: 55f51e85aa61abcb1bc9ab735cd1e863152c46f284a85ff78355306b0698305772b968c1690d7c0c991715afdd433911ac70fce6ce0fde5f133fa27001748ffc
+  metadata.gz: decb847bc6cfd50e816c916d98c74b837d63e9e38a9b49a2d964333a141a8fafd9e55291f83a52739403ad99126884c2a37928567ac9f4f8fdaae3ca9df37692
+  data.tar.gz: 59be5497f4f1a4d439e86b4138be240b1bcabb370bc6d991d79b92aa64e9a40f0243b29f090d4171fd45146593ed27fc2da29ae2a2a72d0f542919e3e5aecf3b

data/lib/moxml/adapter/nokogiri.rb

@@ -30,6 +30,7 @@ module Moxml
           ::Nokogiri::XML::DocumentFragment.parse(processed) do |config|
             config.strict.nonet
             config.recover
+            config.huge
           end.children.to_a
         end
 
@@ -64,12 +65,18 @@ module Moxml
                 config.strict.nonet
                 config.recover unless options[:strict]
                 config.noblanks if options[:noblanks]
+                config.huge
               end
             else
               ::Nokogiri::XML(processed_xml, nil, "UTF-8") do |config|
                 config.strict.nonet
                 config.recover unless options[:strict]
                 config.noblanks if options[:noblanks]
+                # XML_PARSE_HUGE: lifts libxml2's arbitrary limits —
+                # text nodes > 10MB are silently TRUNCATED at 10MB in
+                # recover mode, so large documents lose data without
+                # it (users parse >10MB XML).
+                config.huge
               end
             end
           rescue ::Nokogiri::XML::SyntaxError => e
@@ -89,7 +96,7 @@ module Moxml
         # malformed input never raises.
         def parse_html(html, _options = {}, _context = nil)
           html_string = html.is_a?(IO) || html.is_a?(StringIO) ? html.read : html.to_s
-          native_doc = ::Nokogiri::HTML(html_string)
+          native_doc = ::Nokogiri::HTML(html_string, &:huge)
           Wrappers::Document.new(native_doc, _context || Context.new(:nokogiri))
         end
 

data/lib/moxml/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Moxml
-  VERSION = "0.5.83"
+  VERSION = "0.5.84"
 end

data/spec/moxml/adapter/nokogiri_spec.rb

@@ -36,4 +36,22 @@ RSpec.describe Moxml::Adapter::Nokogiri do
       expect(doc.parse_errors).to all(be_a(String))
     end
   end
+
+  # libxml2 truncates text nodes over XML_MAX_TEXT_LENGTH (10MB) —
+  # silently in recover mode — unless XML_PARSE_HUGE is set. The
+  # adapter always passes huge so large documents keep their data.
+  describe "XML_PARSE_HUGE (large documents)" do
+    let(:ctx) { Moxml.new(:nokogiri) }
+    let(:huge_xml) { "<r>#{'x' * 10_000_001}</r>" }
+
+    it "keeps text nodes over 10MB intact in recover mode" do
+      doc = ctx.parse(huge_xml, strict: false)
+      expect(doc.root.children.first.text.length).to eq(10_000_001)
+    end
+
+    it "keeps text nodes over 10MB intact in strict mode" do
+      doc = ctx.parse(huge_xml)
+      expect(doc.root.children.first.text.length).to eq(10_000_001)
+    end
+  end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: moxml
 version: !ruby/object:Gem::Version
-  version: 0.5.83
+  version: 0.5.84
 platform: ruby
 authors:
 - Ribose Inc.