nokogiri 1.12.5 → 1.14.3

data/ext/nokogiri/extconf.rb

@@ -1,5 +1,8 @@
 # frozen_string_literal: true
-ENV["RC_ARCHS"] = "" if RUBY_PLATFORM =~ /darwin/
+
+# rubocop:disable Style/GlobalVars
+
+ENV["RC_ARCHS"] = "" if RUBY_PLATFORM.include?("darwin")
 
 require "mkmf"
 require "rbconfig"
@@ -12,16 +15,14 @@ PACKAGE_ROOT_DIR = File.expand_path(File.join(File.dirname(__FILE__), "..", ".."
 REQUIRED_LIBXML_VERSION = "2.6.21"
 RECOMMENDED_LIBXML_VERSION = "2.9.3"
 
-# The gem version constraint in the Rakefile is not respected at install time.
-# Keep this version in sync with the one in the Rakefile !
-REQUIRED_MINI_PORTILE_VERSION = "~> 2.6.1"
+REQUIRED_MINI_PORTILE_VERSION = "~> 2.8.0" # keep this version in sync with the one in the gemspec
 REQUIRED_PKG_CONFIG_VERSION = "~> 1.1"
 
 # Keep track of what versions of what libraries we build against
 OTHER_LIBRARY_VERSIONS = {}
 
 NOKOGIRI_HELP_MESSAGE = <<~HELP
-  USAGE: ruby #{$0} [options]
+  USAGE: ruby #{$PROGRAM_NAME} [options]
 
     Flags that are always valid:
 
@@ -175,23 +176,23 @@ def config_system_libraries?
 end
 
 def windows?
-  RbConfig::CONFIG["target_os"] =~ /mingw32|mswin/
+  RbConfig::CONFIG["target_os"].match?(/mingw|mswin/)
 end
 
 def solaris?
-  RbConfig::CONFIG["target_os"] =~ /solaris/
+  RbConfig::CONFIG["target_os"].include?("solaris")
 end
 
 def darwin?
-  RbConfig::CONFIG["target_os"] =~ /darwin/
+  RbConfig::CONFIG["target_os"].include?("darwin")
 end
 
 def openbsd?
-  RbConfig::CONFIG["target_os"] =~ /openbsd/
+  RbConfig::CONFIG["target_os"].include?("openbsd")
 end
 
 def aix?
-  RbConfig::CONFIG["target_os"] =~ /aix/
+  RbConfig::CONFIG["target_os"].include?("aix")
 end
 
 def nix?
@@ -199,7 +200,7 @@ def nix?
 end
 
 def truffle?
-  ::RUBY_ENGINE == "truffleruby"
+  RUBY_ENGINE == "truffleruby"
 end
 
 def concat_flags(*args)
@@ -210,6 +211,28 @@ def local_have_library(lib, func = nil, headers = nil)
   have_library(lib, func, headers) || have_library("lib#{lib}", func, headers)
 end
 
+def zlib_source(version_string)
+  # As of 2022-12, I'm starting to see failed downloads often enough from zlib.net that I want to
+  # change the default to github.
+  if ENV["NOKOGIRI_USE_CANONICAL_ZLIB_SOURCE"]
+    "https://zlib.net/fossils/zlib-#{version_string}.tar.gz"
+  else
+    "https://github.com/madler/zlib/releases/download/v#{version_string}/zlib-#{version_string}.tar.gz"
+  end
+end
+
+def gnome_source
+  # As of 2022-02-20, some mirrors have expired SSL certificates. I'm able to retrieve from my home,
+  # but whatever host is resolved on the github actions workers see an expired cert.
+  #
+  # See https://github.com/sparklemotion/nokogiri/runs/5266206403?check_suite_focus=true
+  if ENV["NOKOGIRI_USE_CANONICAL_GNOME_SOURCE"]
+    "https://download.gnome.org"
+  else
+    "https://mirror.csclub.uwaterloo.ca/gnome" # old reliable
+  end
+end
+
 LOCAL_PACKAGE_RESPONSE = Object.new
 def LOCAL_PACKAGE_RESPONSE.%(package)
   package ? "yes: #{package}" : "no"
@@ -284,17 +307,16 @@ ensure
 end
 
 def abort_could_not_find_library(lib)
-  abort("-----\n#{caller[0]}\n#{lib} is missing. Please locate mkmf.log to investigate how it is failing.\n-----")
+  callers = caller(1..2).join("\n")
+  abort("-----\n#{callers}\n#{lib} is missing. Please locate mkmf.log to investigate how it is failing.\n-----")
 end
 
-def chdir_for_build
+def chdir_for_build(&block)
   # When using rake-compiler-dock on Windows, the underlying Virtualbox shared
   # folders don't support symlinks, but libiconv expects it for a build on
   # Linux. We work around this limitation by using the temp dir for cooking.
-  build_dir = ENV["RCD_HOST_RUBY_PLATFORM"].to_s =~ /mingw|mswin|cygwin/ ? "/tmp" : "."
-  Dir.chdir(build_dir) do
-    yield
-  end
+  build_dir = /mingw|mswin|cygwin/.match?(ENV["RCD_HOST_RUBY_PLATFORM"].to_s) ? "/tmp" : "."
+  Dir.chdir(build_dir, &block)
 end
 
 def sh_export_path(path)
@@ -370,6 +392,7 @@ def iconv_configure_flags
   ["iconv", "opt"].each do |target|
     config = preserving_globals { dir_config(target) }
     next unless config.any? && try_link_iconv("--with-#{target}-* flags") { dir_config(target) }
+
     idirs, ldirs = config.map do |dirs|
       Array(dirs).flat_map do |dir|
         dir.split(File::PATH_SEPARATOR)
@@ -387,8 +410,8 @@ def iconv_configure_flags
     return ["--with-iconv=yes"]
   end
 
-  config = preserving_globals { have_package_configuration("libiconv") }
-  if config && try_link_iconv("pkg-config libiconv") { have_package_configuration("libiconv") }
+  config = preserving_globals { pkg_config("libiconv") }
+  if config && try_link_iconv("pkg-config libiconv") { pkg_config("libiconv") }
     cflags, ldflags, libs = config
 
     return [
@@ -402,9 +425,9 @@ def iconv_configure_flags
   abort_could_not_find_library("libiconv")
 end
 
-def process_recipe(name, version, static_p, cross_p, cacheable_p=true)
+def process_recipe(name, version, static_p, cross_p, cacheable_p = true)
   require "rubygems"
-  gem("mini_portile2", REQUIRED_MINI_PORTILE_VERSION)
+  gem("mini_portile2", REQUIRED_MINI_PORTILE_VERSION) # gemspec is not respected at install time
   require "mini_portile2"
   message("Using mini_portile version #{MiniPortile::VERSION}\n")
 
@@ -413,10 +436,16 @@ def process_recipe(name, version, static_p, cross_p, cacheable_p=true)
   end
 
   MiniPortile.new(name, version).tap do |recipe|
-    recipe.target = File.join(PACKAGE_ROOT_DIR, "ports") if cacheable_p
-    # Prefer host_alias over host in order to use i586-mingw32msvc as
-    # correct compiler prefix for cross build, but use host if not set.
+    def recipe.port_path
+      "#{@target}/#{RUBY_PLATFORM}/#{@name}/#{@version}"
+    end
+
+    # We use 'host' to set compiler prefix for cross-compiling. Prefer host_alias over host. And
+    # prefer i686 (what external dev tools use) to i386 (what ruby's configure.ac emits).
     recipe.host = RbConfig::CONFIG["host_alias"].empty? ? RbConfig::CONFIG["host"] : RbConfig::CONFIG["host_alias"]
+    recipe.host = recipe.host.gsub(/i386/, "i686")
+
+    recipe.target = File.join(PACKAGE_ROOT_DIR, "ports") if cacheable_p
     recipe.configure_options << "--libdir=#{File.join(recipe.path, "lib")}"
 
     yield recipe
@@ -462,7 +491,7 @@ def process_recipe(name, version, static_p, cross_p, cacheable_p=true)
     end
 
     if RbConfig::CONFIG["target_cpu"] == "universal"
-      %w[CFLAGS LDFLAGS].each do |key|
+      ["CFLAGS", "LDFLAGS"].each do |key|
         unless env[key].include?("-arch")
           env[key] = concat_flags(env[key], RbConfig::CONFIG["ARCH_FLAG"])
         end
@@ -473,7 +502,7 @@ def process_recipe(name, version, static_p, cross_p, cacheable_p=true)
       "#{key}=#{value.strip}"
     end
 
-    checkpoint = "#{recipe.target}/#{recipe.name}-#{recipe.version}-#{recipe.host}.installed"
+    checkpoint = "#{recipe.target}/#{recipe.name}-#{recipe.version}-#{RUBY_PLATFORM}.installed"
     if File.exist?(checkpoint) && !recipe.source_directory
       message("Building Nokogiri with a packaged version of #{name}-#{version}.\n")
     else
@@ -487,7 +516,7 @@ def process_recipe(name, version, static_p, cross_p, cacheable_p=true)
         message("The following patches are being applied:\n")
 
         recipe.patch_files.each do |patch|
-          message("  - %s\n" % File.basename(patch))
+          message(format("  - %s\n", File.basename(patch)))
         end
       end
 
@@ -555,6 +584,34 @@ def do_clean
   exit!(0)
 end
 
+# In ruby 3.2, symbol resolution changed on Darwin, to introduce the `-bundle_loader` flag to
+# resolve symbols against the ruby binary.
+#
+# This makes it challenging to build a single extension that works with both a ruby with
+# `--enable-shared` and one with `--disable-shared. To work around that, we choose to add
+# `-flat_namespace` to the link line (later in this file).
+#
+# The `-flat_namespace` line introduces its own behavior change, which is that (similar to on
+# Linux), any symbols in the extension that are exported may now be resolved by shared libraries
+# loaded by the Ruby process. Specifically, that means that libxml2 and libxslt, which are
+# statically linked into the nokogiri bundle, will resolve (at runtime) to a system libxml2 loaded
+# by Ruby on Darwin. And it appears that often Ruby on Darwin does indeed load the system libxml2,
+# and that messes with our assumptions about whether we're running with a patched libxml2 or a
+# vanilla libxml2.
+#
+# We choose to use `-load_hidden` in this case to prevent exporting those symbols from libxml2 and
+# libxslt, which ensures that they will be resolved to the static libraries in the bundle. In other
+# words, when we use `load_hidden`, what happens in the extension stays in the extension.
+#
+# See https://github.com/rake-compiler/rake-compiler-dock/issues/87 for more info.
+#
+# Anyway, this method is the logical bit to tell us when to turn on these workarounds.
+def needs_darwin_linker_hack
+  config_cross_build? &&
+    darwin? &&
+    Gem::Requirement.new("~> 3.2").satisfied_by?(Gem::Version.new(RbConfig::CONFIG["ruby_version"].split("+").first))
+end
+
 #
 #  main
 #
@@ -562,7 +619,7 @@ do_help if arg_config("--help")
 do_clean if arg_config("--clean")
 
 if openbsd? && !config_system_libraries?
-  if %x(#{ENV["CC"] || "/usr/bin/cc"} -v 2>&1) !~ /clang/
+  unless %x(#{ENV["CC"] || "/usr/bin/cc"} -v 2>&1).include?("clang")
     (ENV["CC"] ||= find_executable("egcc")) ||
       abort("Please install gcc 4.9+ from ports using `pkg_add -v gcc`")
   end
@@ -598,6 +655,9 @@ $LIBS = concat_flags($LIBS, ENV["LIBS"])
 # errors/warnings. see #2302
 append_cflags(["-std=c99", "-Wno-declaration-after-statement"])
 
+# gumbo html5 serialization is slower with O3, let's make sure we use O2
+append_cflags("-O2")
+
 # always include debugging information
 append_cflags("-g")
 
@@ -607,8 +667,18 @@ append_cflags("-Winline")
 # good to have no matter what Ruby was compiled with
 append_cflags("-Wmissing-noreturn")
 
+# check integer loss of precision
+if darwin?
+  append_cflags("-Wshorten-64-to-32")
+else
+  append_cflags("-Wconversion -Wno-sign-conversion")
+end
+
 # handle clang variations, see #1101
-append_cflags("-Wno-error=unused-command-line-argument-hard-error-in-future") if darwin?
+if darwin?
+  append_cflags("-Wno-error=unused-command-line-argument-hard-error-in-future")
+  append_cflags("-Wno-unknown-warning-option")
+end
 
 # these tend to be noisy, but on occasion useful during development
 # append_cflags(["-Wcast-qual", "-Wwrite-strings"])
@@ -648,6 +718,10 @@ else
   cross_build_p = config_cross_build?
   message "Cross build is #{cross_build_p ? "enabled" : "disabled"}.\n"
 
+  if needs_darwin_linker_hack
+    append_ldflags("-Wl,-flat_namespace")
+  end
+
   require "yaml"
   dependencies = YAML.load_file(File.join(PACKAGE_ROOT_DIR, "dependencies.yml"))
 
@@ -656,7 +730,7 @@ else
   if cross_build_p || windows?
     zlib_recipe = process_recipe("zlib", dependencies["zlib"]["version"], static_p, cross_build_p) do |recipe|
       recipe.files = [{
-        url: "http://zlib.net/fossils/#{recipe.name}-#{recipe.version}.tar.gz",
+        url: zlib_source(recipe.version),
         sha256: dependencies["zlib"]["sha256"],
       }]
       if windows?
@@ -694,16 +768,16 @@ else
       else
         class << recipe
           def configure
-            cflags = concat_flags(ENV["CFLAGS"], "-fPIC", "-g")
-            execute("configure",
-              ["env", "CHOST=#{host}", "CFLAGS=#{cflags}", "./configure", "--static", configure_prefix])
-          end
-
-          def compile
-            if host =~ /darwin/
-              execute("compile", "make AR=#{host}-libtool")
-            else
-              super
+            env = {}
+            env["CFLAGS"] = concat_flags(ENV["CFLAGS"], "-fPIC", "-g")
+            env["CHOST"] = host
+            execute("configure", ["./configure", "--static", configure_prefix], { env: env })
+            if darwin?
+              # needed as of zlib 1.2.13
+              Dir.chdir(work_path) do
+                makefile = File.read("Makefile").gsub(/^AR=.*$/, "AR=#{host}-libtool")
+                File.open("Makefile", "w") { |m| m.write(makefile) }
+              end
             end
           end
         end
@@ -714,10 +788,13 @@ else
       libiconv_recipe = process_recipe("libiconv", dependencies["libiconv"]["version"], static_p,
         cross_build_p) do |recipe|
         recipe.files = [{
-          url: "http://ftp.gnu.org/pub/gnu/libiconv/#{recipe.name}-#{recipe.version}.tar.gz",
+          url: "https://ftp.gnu.org/pub/gnu/libiconv/#{recipe.name}-#{recipe.version}.tar.gz",
           sha256: dependencies["libiconv"]["sha256"],
         }]
 
+        # The libiconv configure script doesn't accept "arm64" host string but "aarch64"
+        recipe.host = recipe.host.gsub("arm64-apple-darwin", "aarch64-apple-darwin")
+
         cflags = concat_flags(ENV["CFLAGS"], "-O2", "-U_FORTIFY_SOURCE", "-g")
 
         recipe.configure_options += [
@@ -741,12 +818,21 @@ else
       Tools" to open the developer site, download the installer for your OS
       version and run it.
       -----
-      EOM
+    EOM
   end
 
-  unless windows?
-    preserving_globals { local_have_library("z", "gzdopen", "zlib.h") } ||
-      abort("zlib is missing; necessary for building libxml2")
+  if zlib_recipe
+    append_cppflags("-I#{zlib_recipe.path}/include")
+    $LIBPATH = ["#{zlib_recipe.path}/lib"] | $LIBPATH
+    ensure_package_configuration(opt: "zlib", pc: "zlib", lib: "z",
+      headers: "zlib.h", func: "gzdopen")
+  end
+
+  if libiconv_recipe
+    append_cppflags("-I#{libiconv_recipe.path}/include")
+    $LIBPATH = ["#{libiconv_recipe.path}/lib"] | $LIBPATH
+    ensure_package_configuration(opt: "iconv", pc: "iconv", lib: "iconv",
+      headers: "iconv.h", func: "iconv_open")
   end
 
   libxml2_recipe = process_recipe("libxml2", dependencies["libxml2"]["version"], static_p, cross_build_p) do |recipe|
@@ -754,8 +840,9 @@ else
     if source_dir
       recipe.source_directory = source_dir
     else
+      minor_version = Gem::Version.new(recipe.version).segments.take(2).join(".")
       recipe.files = [{
-        url: "http://xmlsoft.org/sources/#{recipe.name}-#{recipe.version}.tar.gz",
+        url: "#{gnome_source}/sources/libxml2/#{minor_version}/#{recipe.name}-#{recipe.version}.tar.xz",
         sha256: dependencies["libxml2"]["sha256"],
       }]
       recipe.patch_files = Dir[File.join(PACKAGE_ROOT_DIR, "patches", "libxml2", "*.patch")].sort
@@ -765,7 +852,6 @@ else
 
     if zlib_recipe
       recipe.configure_options << "--with-zlib=#{zlib_recipe.path}"
-      cflags = concat_flags(cflags, "-I#{zlib_recipe.path}/include")
     end
 
     if libiconv_recipe
@@ -803,8 +889,9 @@ else
     if source_dir
       recipe.source_directory = source_dir
     else
+      minor_version = Gem::Version.new(recipe.version).segments.take(2).join(".")
       recipe.files = [{
-        url: "http://xmlsoft.org/sources/#{recipe.name}-#{recipe.version}.tar.gz",
+        url: "#{gnome_source}/sources/libxslt/#{minor_version}/#{recipe.name}-#{recipe.version}.tar.xz",
         sha256: dependencies["libxslt"]["sha256"],
       }]
       recipe.patch_files = Dir[File.join(PACKAGE_ROOT_DIR, "patches", "libxslt", "*.patch")].sort
@@ -816,6 +903,11 @@ else
       recipe.configure_options += ["RANLIB=/usr/bin/ranlib", "AR=/usr/bin/ar"]
     end
 
+    if windows?
+      cflags = concat_flags(cflags, "-ULIBXSLT_STATIC", "-DIN_LIBXSLT")
+      cflags = concat_flags(cflags, "-ULIBEXSLT_STATIC", "-DIN_LIBEXSLT")
+    end
+
     recipe.configure_options << if source_dir
       "--config-cache"
     else
@@ -834,15 +926,16 @@ else
   append_cppflags("-DNOKOGIRI_PACKAGED_LIBRARIES")
   append_cppflags("-DNOKOGIRI_PRECOMPILED_LIBRARIES") if cross_build_p
 
-  $LIBPATH = ["#{zlib_recipe.path}/lib"] | $LIBPATH if zlib_recipe
-  $LIBPATH = ["#{libiconv_recipe.path}/lib"] | $LIBPATH if libiconv_recipe
-
   $libs = $libs.shellsplit.tap do |libs|
     [libxml2_recipe, libxslt_recipe].each do |recipe|
       libname = recipe.name[/\Alib(.+)\z/, 1]
-      File.join(recipe.path, "bin", "#{libname}-config").tap do |config|
+      config_basename = "#{libname}-config"
+      File.join(recipe.path, "bin", config_basename).tap do |config|
         # call config scripts explicit with 'sh' for compat with Windows
-        $CPPFLAGS = %x(sh #{config} --cflags).strip << " " << $CPPFLAGS
+        cflags = %x(sh #{config} --cflags).strip
+        message("#{config_basename} cflags: #{cflags}\n")
+        $CPPFLAGS = concat_flags(cflags, $CPPFLAGS) # prepend
+
         %x(sh #{config} --libs).strip.shellsplit.each do |arg|
           case arg
           when /\A-L(.+)\z/
@@ -861,7 +954,7 @@ else
       end
 
       patches_string = recipe.patch_files.map { |path| File.basename(path) }.join(" ")
-      append_cppflags(%[-DNOKOGIRI_#{recipe.name.upcase}_PATCHES="\\\"#{patches_string}\\\""])
+      append_cppflags(%[-DNOKOGIRI_#{recipe.name.upcase}_PATCHES="\\"#{patches_string}\\""])
 
       case libname
       when "xml2"
@@ -880,16 +973,17 @@ else
   end.shelljoin
 
   if static_p
+    static_archive_ld_flag = needs_darwin_linker_hack ? ["-load_hidden"] : []
     $libs = $libs.shellsplit.map do |arg|
       case arg
       when "-lxml2"
-        File.join(libxml2_recipe.path, "lib", libflag_to_filename(arg))
+        static_archive_ld_flag + [File.join(libxml2_recipe.path, "lib", libflag_to_filename(arg))]
       when "-lxslt", "-lexslt"
-        File.join(libxslt_recipe.path, "lib", libflag_to_filename(arg))
+        static_archive_ld_flag + [File.join(libxslt_recipe.path, "lib", libflag_to_filename(arg))]
       else
         arg
       end
-    end.shelljoin
+    end.flatten.shelljoin
   end
 
   ensure_func("xmlParseDoc", "libxml/parser.h")
@@ -925,11 +1019,11 @@ libgumbo_recipe = process_recipe("libgumbo", "1.0.0-nokogiri", static_p, cross_b
     end
 
     def compile
-      cflags = concat_flags(ENV["CFLAGS"], "-fPIC", "-g")
+      cflags = concat_flags(ENV["CFLAGS"], "-fPIC", "-O2", "-g")
 
       env = { "CC" => gcc_cmd, "CFLAGS" => cflags }
       if config_cross_build?
-        if host =~ /darwin/
+        if /darwin/.match?(host)
           env["AR"] = "#{host}-libtool"
           env["ARFLAGS"] = "-o"
         else
@@ -945,7 +1039,7 @@ end
 append_cppflags("-I#{File.join(libgumbo_recipe.path, "include")}")
 $libs = $libs + " " + File.join(libgumbo_recipe.path, "lib", "libgumbo.a")
 $LIBPATH = $LIBPATH | [File.join(libgumbo_recipe.path, "lib")]
-ensure_func("gumbo_parse_with_options", "gumbo.h")
+ensure_func("gumbo_parse_with_options", "nokogiri_gumbo.h")
 
 have_func("xmlHasFeature") || abort("xmlHasFeature() is missing.") # introduced in libxml 2.6.21
 have_func("xmlFirstElementChild") # introduced in libxml 2.7.3
@@ -953,11 +1047,11 @@ have_func("xmlRelaxNGSetParserStructuredErrors") # introduced in libxml 2.6.24
 have_func("xmlRelaxNGSetValidStructuredErrors") # introduced in libxml 2.6.21
 have_func("xmlSchemaSetValidStructuredErrors") # introduced in libxml 2.6.23
 have_func("xmlSchemaSetParserStructuredErrors") # introduced in libxml 2.6.23
-
-have_func("vasprintf")
+have_func("rb_gc_location") # introduced in Ruby 2.7
+have_func("rb_category_warning") # introduced in Ruby 3.0
 
 other_library_versions_string = OTHER_LIBRARY_VERSIONS.map { |k, v| [k, v].join(":") }.join(",")
-append_cppflags(%[-DNOKOGIRI_OTHER_LIBRARY_VERSIONS="\\\"#{other_library_versions_string}\\\""])
+append_cppflags(%[-DNOKOGIRI_OTHER_LIBRARY_VERSIONS="\\"#{other_library_versions_string}\\""])
 
 unless config_system_libraries?
   if cross_build_p
@@ -986,3 +1080,5 @@ if config_clean?
     EOF
   end
 end
+
+# rubocop:enable Style/GlobalVars

data/ext/nokogiri/gumbo.c

@@ -23,13 +23,13 @@
 //
 // Processing starts by calling gumbo_parse_with_options. The resulting document tree
 // is then walked, a parallel libxml2 tree is constructed, and the final document is
-// then wrapped using Nokogiri_wrap_xml_document. This approach reduces memory and CPU
+// then wrapped using noko_xml_document_wrap. This approach reduces memory and CPU
 // requirements as Ruby objects are only built when necessary.
 //
 
 #include <nokogiri.h>
 
-#include "gumbo.h"
+#include "nokogiri_gumbo.h"
 
 VALUE cNokogiriHtml5Document;
 
@@ -281,12 +281,12 @@ add_errors(const GumboOutput *output, VALUE rdoc, VALUE input, VALUE url)
       rb_iv_set(syntax_error, "@code", INT2NUM(1));   // XML_ERR_INTERNAL_ERROR
       rb_iv_set(syntax_error, "@level", INT2NUM(2));  // XML_ERR_ERROR
       rb_iv_set(syntax_error, "@file", url);
-      rb_iv_set(syntax_error, "@line", INT2NUM(position.line));
+      rb_iv_set(syntax_error, "@line", SIZET2NUM(position.line));
       rb_iv_set(syntax_error, "@str1", str1);
       rb_iv_set(syntax_error, "@str2", Qnil);
       rb_iv_set(syntax_error, "@str3", Qnil);
       rb_iv_set(syntax_error, "@int1", INT2NUM(0));
-      rb_iv_set(syntax_error, "@column", INT2NUM(position.column));
+      rb_iv_set(syntax_error, "@column", SIZET2NUM(position.column));
       rb_ary_push(rerrors, syntax_error);
     }
     rb_iv_set(rdoc, "@errors", rerrors);
@@ -297,6 +297,7 @@ typedef struct {
   GumboOutput *output;
   VALUE input;
   VALUE url_or_frag;
+  VALUE klass;
   xmlDocPtr doc;
 } ParseArgs;
 
@@ -321,7 +322,7 @@ static VALUE parse_continue(VALUE parse_args);
  *  @!visibility protected
  */
 static VALUE
-parse(VALUE self, VALUE input, VALUE url, VALUE max_attributes, VALUE max_errors, VALUE max_depth)
+parse(VALUE self, VALUE input, VALUE url, VALUE max_attributes, VALUE max_errors, VALUE max_depth, VALUE klass)
 {
   GumboOptions options = kGumboDefaultOptions;
   options.max_attributes = NUM2INT(max_attributes);
@@ -333,6 +334,7 @@ parse(VALUE self, VALUE input, VALUE url, VALUE max_attributes, VALUE max_errors
     .output = output,
     .input = input,
     .url_or_frag = url,
+    .klass = klass,
     .doc = NULL,
   };
 
@@ -357,7 +359,9 @@ parse_continue(VALUE parse_args)
   }
   args->doc = doc; // Make sure doc gets cleaned up if an error is thrown.
   build_tree(doc, (xmlNodePtr)doc, output->document);
-  VALUE rdoc = Nokogiri_wrap_xml_document(cNokogiriHtml5Document, doc);
+  VALUE rdoc = noko_xml_document_wrap(args->klass, doc);
+  rb_iv_set(rdoc, "@url", args->url_or_frag);
+  rb_iv_set(rdoc, "@quirks_mode", INT2NUM(output->document->v.document.doc_type_quirks_mode));
   args->doc = NULL; // The Ruby runtime now owns doc so don't delete it.
   add_errors(output, rdoc, args->input, args->url_or_frag);
   return rdoc;
@@ -401,7 +405,7 @@ static xmlNodePtr
 extract_xml_node(VALUE node)
 {
   xmlNodePtr xml_node;
-  Data_Get_Struct(node, xmlNode, xml_node);
+  Noko_Node_Get_Struct(node, xmlNode, xml_node);
   return xml_node;
 }
 
@@ -498,9 +502,11 @@ error:
     }
 
     // Encoding.
-    if (RSTRING_LEN(tag_name) == 14
+    if (ctx_ns == GUMBO_NAMESPACE_MATHML
+        && RSTRING_LEN(tag_name) == 14
         && !st_strcasecmp(ctx_tag, "annotation-xml")) {
       VALUE enc = rb_funcall(ctx, rb_intern_const("[]"),
+                             1,
                              rb_utf8_str_new_static("encoding", 8));
       if (RTEST(enc)) {
         Check_Type(enc, T_STRING);
@@ -512,8 +518,11 @@ error:
   // Quirks mode.
   VALUE doc = rb_funcall(doc_fragment, rb_intern_const("document"), 0);
   VALUE dtd = rb_funcall(doc, internal_subset, 0);
-  if (NIL_P(dtd)) {
+  VALUE doc_quirks_mode = rb_iv_get(doc, "@quirks_mode");
+  if (NIL_P(ctx) || NIL_P(doc_quirks_mode)) {
     quirks_mode = GUMBO_DOCTYPE_NO_QUIRKS;
+  } else if (NIL_P(dtd)) {
+    quirks_mode = GUMBO_DOCTYPE_QUIRKS;
   } else {
     VALUE dtd_name = rb_funcall(dtd, name, 0);
     VALUE pubid = rb_funcall(dtd, rb_intern_const("external_id"), 0);
@@ -560,13 +569,14 @@ fragment_continue(VALUE parse_args)
   args->doc = NULL; // The Ruby runtime owns doc so make sure we don't delete it.
   xmlNodePtr xml_frag = extract_xml_node(doc_fragment);
   build_tree(xml_doc, xml_frag, output->root);
+  rb_iv_set(doc_fragment, "@quirks_mode", INT2NUM(output->document->v.document.doc_type_quirks_mode));
   add_errors(output, doc_fragment, args->input, rb_utf8_str_new_static("#fragment", 9));
   return Qnil;
 }
 
 // Initialize the Nokogumbo class and fetch constants we will use later.
 void
-noko_init_gumbo()
+noko_init_gumbo(void)
 {
   // Class constants.
   cNokogiriHtml5Document = rb_define_class_under(mNokogiriHtml5, "Document", cNokogiriHtml4Document);
@@ -577,7 +587,7 @@ noko_init_gumbo()
   parent = rb_intern_const("parent");
 
   // Define Nokogumbo module with parse and fragment methods.
-  rb_define_singleton_method(mNokogiriGumbo, "parse", parse, 5);
+  rb_define_singleton_method(mNokogiriGumbo, "parse", parse, 6);
   rb_define_singleton_method(mNokogiriGumbo, "fragment", fragment, 6);
 }
 

data/ext/nokogiri/html4_document.c

@@ -146,11 +146,11 @@ rb_html_document_type(VALUE self)
 {
   htmlDocPtr doc;
   Data_Get_Struct(self, xmlDoc, doc);
-  return INT2NUM((long)doc->type);
+  return INT2NUM(doc->type);
 }
 
 void
-noko_init_html_document()
+noko_init_html_document(void)
 {
   assert(cNokogiriXmlDocument);
   cNokogiriHtml4Document = rb_define_class_under(mNokogiriHtml4, "Document", cNokogiriXmlDocument);

data/ext/nokogiri/html4_element_description.c

@@ -270,7 +270,7 @@ get_description(VALUE klass, VALUE tag_name)
 }
 
 void
-noko_init_html_element_description()
+noko_init_html_element_description(void)
 {
   cNokogiriHtml4ElementDescription = rb_define_class_under(mNokogiriHtml4, "ElementDescription", rb_cObject);
 

data/ext/nokogiri/html4_entity_lookup.c

@@ -20,7 +20,7 @@ get(VALUE _, VALUE rb_entity_name)
     return Qnil;
   }
 
-  rb_constructor_args[0] = INT2NUM((long)c_entity_desc->value);
+  rb_constructor_args[0] = UINT2NUM(c_entity_desc->value);
   rb_constructor_args[1] = NOKOGIRI_STR_NEW2(c_entity_desc->name);
   rb_constructor_args[2] = NOKOGIRI_STR_NEW2(c_entity_desc->desc);
 
@@ -29,7 +29,7 @@ get(VALUE _, VALUE rb_entity_name)
 }
 
 void
-noko_init_html_entity_lookup()
+noko_init_html_entity_lookup(void)
 {
   cNokogiriHtml4EntityLookup = rb_define_class_under(mNokogiriHtml4, "EntityLookup", rb_cObject);