nokogiri 1.14.2 → 1.16.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: b926bf2ba1c28516c2dfc541d4b82e0f92ef3c585d6843334d7f1d44fe33abcb
-  data.tar.gz: ec0ff5c75bc44ed7fe85761cfa00b28e92a03afdcd525356d31bd74060a2dbe5
+  metadata.gz: f268b2edaa2782a3e9a64e671e36d88149ef5dc23fcf3967a8a54c4475ed7b85
+  data.tar.gz: b49fbf8708f4733e34c211935637ad5999d3950b89aca2da71899c71eaa14901
 SHA512:
-  metadata.gz: 11d8784c27eb523c7cba2c67efd1321c10253ce9624217b1b0383d246354dd8328466474c7968a7844b8ca01ca8e450e7d35bfee56d917b381949e721b435ce4
-  data.tar.gz: '08db49457bb62572e68f5ec8b1eb29da18a9a632c6ef6ebb4ac9e193a29c0beac253d8aa2ed96a35513e70eb7a495344742d70e6b8ad81126b52e07ed400569d'
+  metadata.gz: e2f096bf7e99c453450c2a9d78c7274a3472d29ea8e8aeeff2dfb6976b5b009a8bfd5e673c7eb1a2575d820df26a830ce4e4043d33fb4c0239bf8098bdb98188
+  data.tar.gz: aa7658f376b0278150550d5426ddbe11469be8624d6fc732cf035d0ed00736e177e2848182c9077a9cc724ae6efe92666e8f3664b71282c4e74e79ad48c51373

data/Gemfile

@@ -5,34 +5,38 @@ source "https://rubygems.org"
 gemspec
 
 group :development do
+  # ruby 3.4.0-dev removed some gems from the default set
+  #
+  # TODO: we should be able to remove these as our gem dependencies sort it out and we pull them in
+  # transitively.
+  gem "mutex_m"
+
   # bootstrapping
   gem "bundler", "~> 2.3"
-  gem "rake", "= 13.0.6"
+  gem "rake", "13.1.0"
 
   # building extensions
-  gem "rake-compiler", "= 1.2.1"
-  gem "rake-compiler-dock", "= 1.3.0"
-
-  # documentation
-  gem "hoe-markdown", "= 1.4.0"
+  gem "rake-compiler", "1.2.6"
+  gem "rake-compiler-dock", "1.4.0"
 
   # parser generator
   gem "rexical", "= 1.0.7"
 
   # tests
-  gem "minitest", "5.17.0"
-  gem "minitest-reporters", "= 1.5.0"
-  gem "ruby_memcheck", "1.2.0" unless RUBY_PLATFORM == "java"
-  gem "simplecov", "= 0.21.2"
+  gem "minitest", "5.21.2"
+  gem "minitest-parallel_fork", "2.0.0"
+  gem "ruby_memcheck", "2.3.0"
   gem "rubyzip", "~> 2.3.2"
+  gem "simplecov", "= 0.21.2"
 
   # rubocop
   if Gem::Requirement.new("~> 3.0").satisfied_by?(Gem::Version.new(RUBY_VERSION))
-    gem "rubocop", "1.44.1"
-    gem "rubocop-minitest", "0.27.0"
-    gem "rubocop-performance", "1.15.2"
+    gem "rubocop", "1.60.2"
+    gem "rubocop-minitest", "0.34.5"
+    gem "rubocop-packaging", "0.5.2"
+    gem "rubocop-performance", "1.20.2"
     gem "rubocop-rake", "= 0.6.0"
-    gem "rubocop-shopify", "2.10.1"
+    gem "rubocop-shopify", "2.14.0"
   end
 end
 
@@ -40,5 +44,5 @@ end
 # `bundle config set --local without rdoc`
 # Then re-run `bundle install`.
 group :rdoc do
-  gem "rdoc", "6.5.0"
+  gem "rdoc", "6.6.2"
 end

data/README.md

@@ -55,6 +55,7 @@ There are a few ways to ask exploratory questions:
 
 - The Nokogiri mailing list is active at https://groups.google.com/group/nokogiri-talk
 - Open an issue using the "Help Request" template at https://github.com/sparklemotion/nokogiri/issues
+- Open a discussion at https://github.com/sparklemotion/nokogiri/discussions
 
 Please do not mail the maintainers at their personal addresses.
 
@@ -90,6 +91,8 @@ We bump `Major.Minor.Patch` versions following this guidance:
 - Updating packaged libraries for non-security-related reasons.
 - Dropping support for EOLed Ruby versions. [Some folks find this objectionable](https://github.com/sparklemotion/nokogiri/issues/1568), but [SemVer says this is OK if the public API hasn't changed](https://semver.org/#what-should-i-do-if-i-update-my-own-dependencies-without-changing-the-public-api).
 - Backwards-incompatible changes to internal or private methods and constants. These are detailed in the "Changes" section of each changelog entry.
+- Removal of deprecated methods or parameters, after a generous transition period; usually when those methods or parameters are rarely-used or dangerous to the user. Essentially, removals that do not justify a major version bump.
+
 
 `Patch`:
 
@@ -111,7 +114,7 @@ You can help sponsor the maintainers of this software through one of these organ
 
 Requirements:
 
-- Ruby >= 2.7
+- Ruby >= 3.0
 - JRuby >= 9.4.0.0
 
 

data/dependencies.yml

@@ -1,16 +1,17 @@
+---
 libxml2:
-  version: "2.10.3"
-  sha256: "5d2cc3d78bec3dbe212a9d7fa629ada25a7da928af432c93060ff5c17ee28a9c"
-  # sha-256 hash provided in https://download.gnome.org/sources/libxml2/2.10/libxml2-2.10.3.sha256sum
+  version: "2.12.7"
+  sha256: "24ae78ff1363a973e6d8beba941a7945da2ac056e19b53956aeb6927fd6cfb56"
+  # sha-256 hash provided in https://download.gnome.org/sources/libxml2/2.12/libxml2-2.12.7.sha256sum
 
 libxslt:
-  version: "1.1.37"
-  sha256: "3a4b27dc8027ccd6146725950336f1ec520928f320f144eb5fa7990ae6123ab4"
-  # sha-256 hash provided in https://download.gnome.org/sources/libxslt/1.1/libxslt-1.1.37.sha256sum
+  version: "1.1.39"
+  sha256: "2a20ad621148339b0759c4d4e96719362dee64c9a096dbba625ba053846349f0"
+  # sha-256 hash provided in https://download.gnome.org/sources/libxslt/1.1/libxslt-1.1.39.sha256sum
 
 zlib:
-  version: "1.2.13"
-  sha256: "b3a24de97a8fdbc835b9833169501030b8977031bcb54b3b3ac13740f846ab30"
+  version: "1.3.1"
+  sha256: "9a93b2b7dfdac77ceba5a558a580e74667dd6fede4585b91eefb60f03b72df23"
   # SHA-256 hash provided on http://zlib.net/
 
 libiconv:

data/ext/nokogiri/extconf.rb

@@ -15,7 +15,7 @@ PACKAGE_ROOT_DIR = File.expand_path(File.join(File.dirname(__FILE__), "..", ".."
 REQUIRED_LIBXML_VERSION = "2.6.21"
 RECOMMENDED_LIBXML_VERSION = "2.9.3"
 
-REQUIRED_MINI_PORTILE_VERSION = "~> 2.8.0" # keep this version in sync with the one in the gemspec
+REQUIRED_MINI_PORTILE_VERSION = "~> 2.8.2" # 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
@@ -229,7 +229,7 @@ def gnome_source
   if ENV["NOKOGIRI_USE_CANONICAL_GNOME_SOURCE"]
     "https://download.gnome.org"
   else
-    "https://mirror.csclub.uwaterloo.ca/gnome" # old reliable
+    "https://muug.ca/mirror/gnome" # old reliable
   end
 end
 
@@ -372,7 +372,7 @@ def try_link_iconv(using = nil)
       preserving_globals do
         yield if block_given?
 
-        try_link(<<~'SRC', opt)
+        try_link(<<~SRC, opt)
           #include <stdlib.h>
           #include <iconv.h>
           int main(void)
@@ -443,7 +443,7 @@ def process_recipe(name, version, static_p, cross_p, cacheable_p = true)
     # 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.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")}"
@@ -651,8 +651,7 @@ append_cppflags(ENV["CPPFLAGS"].split) unless ENV["CPPFLAGS"].nil?
 append_ldflags(ENV["LDFLAGS"].split) unless ENV["LDFLAGS"].nil?
 $LIBS = concat_flags($LIBS, ENV["LIBS"])
 
-# nokogumbo code uses C90/C99 features, let's make sure older compilers won't give
-# errors/warnings. see #2302
+# libgumbo uses C90/C99 features, see #2302
 append_cflags(["-std=c99", "-Wno-declaration-after-statement"])
 
 # gumbo html5 serialization is slower with O3, let's make sure we use O2
@@ -695,14 +694,34 @@ append_cppflags(' "-Idummypath"') if windows?
 
 if config_system_libraries?
   message "Building nokogiri using system libraries.\n"
-  ensure_package_configuration(opt: "zlib", pc: "zlib", lib: "z",
-    headers: "zlib.h", func: "gzdopen")
-  ensure_package_configuration(opt: "xml2", pc: "libxml-2.0", lib: "xml2",
-    headers: "libxml/parser.h", func: "xmlParseDoc")
-  ensure_package_configuration(opt: "xslt", pc: "libxslt", lib: "xslt",
-    headers: "libxslt/xslt.h", func: "xsltParseStylesheetDoc")
-  ensure_package_configuration(opt: "exslt", pc: "libexslt", lib: "exslt",
-    headers: "libexslt/exslt.h", func: "exsltFuncRegister")
+  ensure_package_configuration(
+    opt: "zlib",
+    pc: "zlib",
+    lib: "z",
+    headers: "zlib.h",
+    func: "gzdopen",
+  )
+  ensure_package_configuration(
+    opt: "xml2",
+    pc: "libxml-2.0",
+    lib: "xml2",
+    headers: "libxml/parser.h",
+    func: "xmlParseDoc",
+  )
+  ensure_package_configuration(
+    opt: "xslt",
+    pc: "libxslt",
+    lib: "xslt",
+    headers: "libxslt/xslt.h",
+    func: "xsltParseStylesheetDoc",
+  )
+  ensure_package_configuration(
+    opt: "exslt",
+    pc: "libexslt",
+    lib: "exslt",
+    headers: "libexslt/exslt.h",
+    func: "exsltFuncRegister",
+  )
 
   have_libxml_headers?(REQUIRED_LIBXML_VERSION) ||
     abort("ERROR: libxml2 version #{REQUIRED_LIBXML_VERSION} or later is required!")
@@ -785,8 +804,12 @@ else
     end
 
     unless nix?
-      libiconv_recipe = process_recipe("libiconv", dependencies["libiconv"]["version"], static_p,
-        cross_build_p) do |recipe|
+      libiconv_recipe = process_recipe(
+        "libiconv",
+        dependencies["libiconv"]["version"],
+        static_p,
+        cross_build_p,
+      ) do |recipe|
         recipe.files = [{
           url: "https://ftp.gnu.org/pub/gnu/libiconv/#{recipe.name}-#{recipe.version}.tar.gz",
           sha256: dependencies["libiconv"]["sha256"],
@@ -824,15 +847,25 @@ else
   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")
+    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")
+    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|
@@ -848,8 +881,13 @@ else
       recipe.patch_files = Dir[File.join(PACKAGE_ROOT_DIR, "patches", "libxml2", "*.patch")].sort
     end
 
+    cppflags = concat_flags(ENV["CPPFLAGS"])
     cflags = concat_flags(ENV["CFLAGS"], "-O2", "-U_FORTIFY_SOURCE", "-g")
 
+    if cross_build_p
+      cppflags = concat_flags(cppflags, "-DNOKOGIRI_PRECOMPILED_LIBRARIES")
+    end
+
     if zlib_recipe
       recipe.configure_options << "--with-zlib=#{zlib_recipe.path}"
     end
@@ -880,6 +918,8 @@ else
       "--with-c14n",
       "--with-debug",
       "--with-threads",
+      "--without-tls", # see https://github.com/sparklemotion/nokogiri/issues/3031
+      "CPPFLAGS=#{cppflags}",
       "CFLAGS=#{cflags}",
     ]
   end
@@ -1023,7 +1063,7 @@ libgumbo_recipe = process_recipe("libgumbo", "1.0.0-nokogiri", static_p, cross_b
 
       env = { "CC" => gcc_cmd, "CFLAGS" => cflags }
       if config_cross_build?
-        if /darwin/.match?(host)
+        if host.include?("darwin")
           env["AR"] = "#{host}-libtool"
           env["ARFLAGS"] = "-o"
         else
@@ -1047,7 +1087,6 @@ 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("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(",")
@@ -1057,12 +1096,16 @@ unless config_system_libraries?
   if cross_build_p
     # When precompiling native gems, copy packaged libraries' headers to ext/nokogiri/include
     # These are packaged up by the cross-compiling callback in the ExtensionTask
-    copy_packaged_libraries_headers(to_path: File.join(PACKAGE_ROOT_DIR, "ext/nokogiri/include"),
-      from_recipes: [libxml2_recipe, libxslt_recipe])
+    copy_packaged_libraries_headers(
+      to_path: File.join(PACKAGE_ROOT_DIR, "ext/nokogiri/include"),
+      from_recipes: [libxml2_recipe, libxslt_recipe],
+    )
   else
     # When compiling during installation, install packaged libraries' header files into ext/nokogiri/include
-    copy_packaged_libraries_headers(to_path: "include",
-      from_recipes: [libxml2_recipe, libxslt_recipe])
+    copy_packaged_libraries_headers(
+      to_path: "include",
+      from_recipes: [libxml2_recipe, libxslt_recipe],
+    )
     $INSTALLFILES << ["include/**/*.h", "$(rubylibdir)"]
   end
 end

data/ext/nokogiri/html4_document.c

@@ -144,8 +144,7 @@ rb_html_document_s_read_memory(VALUE klass, VALUE rb_html, VALUE rb_url, VALUE r
 static VALUE
 rb_html_document_type(VALUE self)
 {
-  htmlDocPtr doc;
-  Data_Get_Struct(self, xmlDoc, doc);
+  htmlDocPtr doc = noko_xml_document_unwrap(self);
   return INT2NUM(doc->type);
 }
 

data/ext/nokogiri/html4_element_description.c

@@ -1,5 +1,10 @@
 #include <nokogiri.h>
 
+static const rb_data_type_t html4_element_description_type = {
+  .wrap_struct_name = "Nokogiri::HTML4::ElementDescription",
+  .flags = RUBY_TYPED_FREE_IMMEDIATELY | RUBY_TYPED_WB_PROTECTED,
+};
+
 VALUE cNokogiriHtml4ElementDescription ;
 
 /*
@@ -15,7 +20,7 @@ required_attributes(VALUE self)
   VALUE list;
   int i;
 
-  Data_Get_Struct(self, htmlElemDesc, description);
+  TypedData_Get_Struct(self, htmlElemDesc, &html4_element_description_type, description);
 
   list = rb_ary_new();
 
@@ -41,7 +46,7 @@ deprecated_attributes(VALUE self)
   VALUE list;
   int i;
 
-  Data_Get_Struct(self, htmlElemDesc, description);
+  TypedData_Get_Struct(self, htmlElemDesc, &html4_element_description_type, description);
 
   list = rb_ary_new();
 
@@ -67,7 +72,7 @@ optional_attributes(VALUE self)
   VALUE list;
   int i;
 
-  Data_Get_Struct(self, htmlElemDesc, description);
+  TypedData_Get_Struct(self, htmlElemDesc, &html4_element_description_type, description);
 
   list = rb_ary_new();
 
@@ -90,7 +95,7 @@ static VALUE
 default_sub_element(VALUE self)
 {
   const htmlElemDesc *description;
-  Data_Get_Struct(self, htmlElemDesc, description);
+  TypedData_Get_Struct(self, htmlElemDesc, &html4_element_description_type, description);
 
   if (description->defaultsubelt) {
     return NOKOGIRI_STR_NEW2(description->defaultsubelt);
@@ -112,7 +117,7 @@ sub_elements(VALUE self)
   VALUE list;
   int i;
 
-  Data_Get_Struct(self, htmlElemDesc, description);
+  TypedData_Get_Struct(self, htmlElemDesc, &html4_element_description_type, description);
 
   list = rb_ary_new();
 
@@ -135,7 +140,7 @@ static VALUE
 description(VALUE self)
 {
   const htmlElemDesc *description;
-  Data_Get_Struct(self, htmlElemDesc, description);
+  TypedData_Get_Struct(self, htmlElemDesc, &html4_element_description_type, description);
 
   return NOKOGIRI_STR_NEW2(description->desc);
 }
@@ -150,7 +155,7 @@ static VALUE
 inline_eh(VALUE self)
 {
   const htmlElemDesc *description;
-  Data_Get_Struct(self, htmlElemDesc, description);
+  TypedData_Get_Struct(self, htmlElemDesc, &html4_element_description_type, description);
 
   if (description->isinline) { return Qtrue; }
   return Qfalse;
@@ -166,7 +171,7 @@ static VALUE
 deprecated_eh(VALUE self)
 {
   const htmlElemDesc *description;
-  Data_Get_Struct(self, htmlElemDesc, description);
+  TypedData_Get_Struct(self, htmlElemDesc, &html4_element_description_type, description);
 
   if (description->depr) { return Qtrue; }
   return Qfalse;
@@ -182,7 +187,7 @@ static VALUE
 empty_eh(VALUE self)
 {
   const htmlElemDesc *description;
-  Data_Get_Struct(self, htmlElemDesc, description);
+  TypedData_Get_Struct(self, htmlElemDesc, &html4_element_description_type, description);
 
   if (description->empty) { return Qtrue; }
   return Qfalse;
@@ -198,7 +203,7 @@ static VALUE
 save_end_tag_eh(VALUE self)
 {
   const htmlElemDesc *description;
-  Data_Get_Struct(self, htmlElemDesc, description);
+  TypedData_Get_Struct(self, htmlElemDesc, &html4_element_description_type, description);
 
   if (description->saveEndTag) { return Qtrue; }
   return Qfalse;
@@ -214,7 +219,7 @@ static VALUE
 implied_end_tag_eh(VALUE self)
 {
   const htmlElemDesc *description;
-  Data_Get_Struct(self, htmlElemDesc, description);
+  TypedData_Get_Struct(self, htmlElemDesc, &html4_element_description_type, description);
 
   if (description->endTag) { return Qtrue; }
   return Qfalse;
@@ -230,7 +235,7 @@ static VALUE
 implied_start_tag_eh(VALUE self)
 {
   const htmlElemDesc *description;
-  Data_Get_Struct(self, htmlElemDesc, description);
+  TypedData_Get_Struct(self, htmlElemDesc, &html4_element_description_type, description);
 
   if (description->startTag) { return Qtrue; }
   return Qfalse;
@@ -246,7 +251,7 @@ static VALUE
 name(VALUE self)
 {
   const htmlElemDesc *description;
-  Data_Get_Struct(self, htmlElemDesc, description);
+  TypedData_Get_Struct(self, htmlElemDesc, &html4_element_description_type, description);
 
   if (NULL == description->name) { return Qnil; }
   return NOKOGIRI_STR_NEW2(description->name);
@@ -266,7 +271,7 @@ get_description(VALUE klass, VALUE tag_name)
                                     );
 
   if (NULL == description) { return Qnil; }
-  return Data_Wrap_Struct(klass, 0, 0, DISCARD_CONST_QUAL(void *, description));
+  return TypedData_Wrap_Struct(klass, &html4_element_description_type, DISCARD_CONST_QUAL(void *, description));
 }
 
 void

data/ext/nokogiri/html4_sax_parser_context.c

@@ -2,13 +2,6 @@
 
 VALUE cNokogiriHtml4SaxParserContext ;
 
-static void
-deallocate(xmlParserCtxtPtr ctxt)
-{
-  ctxt->sax = NULL;
-  htmlFreeParserCtxt(ctxt);
-}
-
 static VALUE
 parse_memory(VALUE klass, VALUE data, VALUE encoding)
 {
@@ -38,7 +31,7 @@ parse_memory(VALUE klass, VALUE data, VALUE encoding)
     }
   }
 
-  return Data_Wrap_Struct(klass, NULL, deallocate, ctxt);
+  return noko_xml_sax_parser_context_wrap(klass, ctxt);
 }
 
 static VALUE
@@ -48,7 +41,13 @@ parse_file(VALUE klass, VALUE filename, VALUE encoding)
                              StringValueCStr(filename),
                              StringValueCStr(encoding)
                            );
-  return Data_Wrap_Struct(klass, NULL, deallocate, ctxt);
+
+  if (ctxt->sax) {
+    xmlFree(ctxt->sax);
+    ctxt->sax = NULL;
+  }
+
+  return noko_xml_sax_parser_context_wrap(klass, ctxt);
 }
 
 static VALUE
@@ -82,13 +81,8 @@ parse_with(VALUE self, VALUE sax_handler)
     rb_raise(rb_eArgError, "argument must be a Nokogiri::XML::SAX::Parser");
   }
 
-  Data_Get_Struct(self, htmlParserCtxt, ctxt);
-  Data_Get_Struct(sax_handler, htmlSAXHandler, sax);
-
-  /* Free the sax handler since we'll assign our own */
-  if (ctxt->sax && ctxt->sax != (xmlSAXHandlerPtr)&xmlDefaultSAXHandler) {
-    xmlFree(ctxt->sax);
-  }
+  ctxt = noko_xml_sax_parser_context_unwrap(self);
+  sax = noko_sax_handler_unwrap(sax_handler);
 
   ctxt->sax = sax;
   ctxt->userData = (void *)NOKOGIRI_SAX_TUPLE_NEW(ctxt, sax_handler);

data/ext/nokogiri/html4_sax_push_parser.c

@@ -17,7 +17,7 @@ native_write(VALUE self, VALUE _chunk, VALUE _last_chunk)
   int status = 0;
   libxmlStructuredErrorHandlerState handler_state;
 
-  Data_Get_Struct(self, xmlParserCtxt, ctx);
+  ctx = noko_xml_sax_push_parser_unwrap(self);
 
   if (Qnil != _chunk) {
     chunk = StringValuePtr(_chunk);
@@ -32,7 +32,7 @@ native_write(VALUE self, VALUE _chunk, VALUE _last_chunk)
 
   if ((status != 0) && !(ctx->options & XML_PARSE_RECOVER)) {
     // TODO: there appear to be no tests for this block
-    xmlErrorPtr e = xmlCtxtGetLastError(ctx);
+    xmlErrorConstPtr e = xmlCtxtGetLastError(ctx);
     Nokogiri_error_raise(NULL, e);
   }
 
@@ -54,7 +54,7 @@ initialize_native(VALUE self, VALUE _xml_sax, VALUE _filename,
   htmlParserCtxtPtr ctx;
   xmlCharEncoding enc = XML_CHAR_ENCODING_NONE;
 
-  Data_Get_Struct(_xml_sax, xmlSAXHandler, sax);
+  sax = noko_sax_handler_unwrap(_xml_sax);
 
   if (_filename != Qnil) { filename = StringValueCStr(_filename); }
 

data/ext/nokogiri/nokogiri.c

@@ -137,6 +137,48 @@ noko_io_close(void *io)
 }
 
 
+#if defined(_WIN32) && !defined(NOKOGIRI_PACKAGED_LIBRARIES)
+#  define NOKOGIRI_WINDOWS_DLLS 1
+#else
+#  define NOKOGIRI_WINDOWS_DLLS 0
+#endif
+
+//
+//   |      dlls || true    | false   |
+//   | nlmm      ||         |         |
+//   |-----------++---------+---------|
+//   | NULL      || default | ruby    |
+//   | "random"  || default | ruby    |
+//   | "ruby"    || ruby    | ruby    |
+//   | "default" || default | default |
+//
+//   We choose *not* to use Ruby's memory management functions with windows DLLs because of this
+//   issue: https://github.com/sparklemotion/nokogiri/issues/2241
+//
+static void
+set_libxml_memory_management(void)
+{
+  const char *nlmm = getenv("NOKOGIRI_LIBXML_MEMORY_MANAGEMENT");
+  if (nlmm) {
+    if (strcmp(nlmm, "default") == 0) {
+      goto libxml_uses_default_memory_management;
+    } else if (strcmp(nlmm, "ruby") == 0) {
+      goto libxml_uses_ruby_memory_management;
+    }
+  }
+  if (NOKOGIRI_WINDOWS_DLLS) {
+libxml_uses_default_memory_management:
+    rb_const_set(mNokogiri, rb_intern("LIBXML_MEMORY_MANAGEMENT"), NOKOGIRI_STR_NEW2("default"));
+    return;
+  } else {
+libxml_uses_ruby_memory_management:
+    rb_const_set(mNokogiri, rb_intern("LIBXML_MEMORY_MANAGEMENT"), NOKOGIRI_STR_NEW2("ruby"));
+    xmlMemSetup((xmlFreeFunc)ruby_xfree, (xmlMallocFunc)ruby_xmalloc, (xmlReallocFunc)ruby_xrealloc, ruby_strdup);
+    return;
+  }
+}
+
+
 void
 Init_nokogiri(void)
 {
@@ -150,6 +192,10 @@ Init_nokogiri(void)
   mNokogiriXmlXpath = rb_define_module_under(mNokogiriXml, "XPath");
   mNokogiriXslt     = rb_define_module_under(mNokogiri, "XSLT");
 
+  set_libxml_memory_management(); /* must be before any function calls that might invoke xmlInitParser() */
+  xmlInitParser();
+  exsltRegisterAll();
+
   rb_const_set(mNokogiri, rb_intern("LIBXML_COMPILED_VERSION"), NOKOGIRI_STR_NEW2(LIBXML_DOTTED_VERSION));
   rb_const_set(mNokogiri, rb_intern("LIBXML_LOADED_VERSION"), NOKOGIRI_STR_NEW2(xmlParserVersion));
 
@@ -182,30 +228,6 @@ Init_nokogiri(void)
   rb_const_set(mNokogiri, rb_intern("OTHER_LIBRARY_VERSIONS"), NOKOGIRI_STR_NEW2(NOKOGIRI_OTHER_LIBRARY_VERSIONS));
 #endif
 
-#if defined(_WIN32) && !defined(NOKOGIRI_PACKAGED_LIBRARIES)
-  /*
-   *  We choose *not* to do use Ruby's memory management functions with windows DLLs because of this
-   *  issue in libxml 2.9.12:
-   *
-   *    https://github.com/sparklemotion/nokogiri/issues/2241
-   *
-   *  If the atexit() issue gets fixed in a future version of libxml2, then we may be able to skip
-   *  this config only for the specific libxml2 versions 2.9.12.
-   *
-   *  Alternatively, now that Ruby has a generational GC, it might be OK to let libxml2 use its
-   *  default memory management functions (recall that this config was introduced to reduce memory
-   *  bloat and allow Ruby to GC more often); but we should *really* test with production workloads
-   *  before making that kind of a potentially-invasive change.
-   */
-  rb_const_set(mNokogiri, rb_intern("LIBXML_MEMORY_MANAGEMENT"), NOKOGIRI_STR_NEW2("default"));
-#else
-  rb_const_set(mNokogiri, rb_intern("LIBXML_MEMORY_MANAGEMENT"), NOKOGIRI_STR_NEW2("ruby"));
-  xmlMemSetup((xmlFreeFunc)ruby_xfree, (xmlMallocFunc)ruby_xmalloc, (xmlReallocFunc)ruby_xrealloc, ruby_strdup);
-#endif
-
-  xmlInitParser();
-  exsltRegisterAll();
-
   if (xsltExtModuleFunctionLookup((const xmlChar *)"date-time", EXSLT_DATE_NAMESPACE)) {
     rb_const_set(mNokogiri, rb_intern("LIBXSLT_DATETIME_ENABLED"), Qtrue);
   } else {