diff --git a/ext/nokogiri/extconf.rb b/ext/nokogiri/extconf.rb
index 9f9ee8dd0a..3f2de4ddfc 100644
--- a/ext/nokogiri/extconf.rb
+++ b/ext/nokogiri/extconf.rb
@@ -1116,6 +1116,8 @@ def compile
   ensure_func("gumbo_parse_with_options", "nokogiri_gumbo.h")
 end
 
+have_func("xmlCtxtSetOptions") # introduced in libxml2 2.13.0
+have_func("xmlCtxtGetOptions") # introduced in libxml2 2.14.0
 have_func("rb_category_warning") # introduced in Ruby 3.0
 
 other_library_versions_string = OTHER_LIBRARY_VERSIONS.map { |k, v| [k, v].join(":") }.join(",")
diff --git a/ext/nokogiri/html4_sax_parser.c b/ext/nokogiri/html4_sax_parser.c
new file mode 100644
index 0000000000..9dba3eb8d4
--- /dev/null
+++ b/ext/nokogiri/html4_sax_parser.c
@@ -0,0 +1,36 @@
+#include <nokogiri.h>
+
+VALUE cNokogiriHtml4SaxParser;
+
+static ID id_start_document;
+
+static void
+noko_html4_sax_parser_start_document(void *ctx)
+{
+  VALUE self = NOKOGIRI_SAX_SELF(ctx);
+  VALUE doc = rb_iv_get(self, "@document");
+
+  rb_funcall(doc, id_start_document, 0);
+}
+
+static VALUE
+noko_html4_sax_parser_initialize(VALUE self)
+{
+  xmlSAXHandlerPtr handler = noko_xml_sax_parser_unwrap(self);
+
+  rb_call_super(0, NULL);
+
+  handler->startDocument = noko_html4_sax_parser_start_document;
+
+  return self;
+}
+
+void
+noko_init_html4_sax_parser(void)
+{
+  cNokogiriHtml4SaxParser = rb_define_class_under(mNokogiriHtml4Sax, "Parser", cNokogiriXmlSaxParser);
+
+  rb_define_private_method(cNokogiriHtml4SaxParser, "initialize_native", noko_html4_sax_parser_initialize, 0);
+
+  id_start_document = rb_intern("start_document");
+}
diff --git a/ext/nokogiri/html4_sax_parser_context.c b/ext/nokogiri/html4_sax_parser_context.c
index 4a5293d0ec..38f102590a 100644
--- a/ext/nokogiri/html4_sax_parser_context.c
+++ b/ext/nokogiri/html4_sax_parser_context.c
@@ -82,7 +82,7 @@ parse_with(VALUE self, VALUE sax_handler)
   }
 
   ctxt = noko_xml_sax_parser_context_unwrap(self);
-  sax = noko_sax_handler_unwrap(sax_handler);
+  sax = noko_xml_sax_parser_unwrap(sax_handler);
 
   ctxt->sax = sax;
   ctxt->userData = (void *)NOKOGIRI_SAX_TUPLE_NEW(ctxt, sax_handler);
diff --git a/ext/nokogiri/html4_sax_push_parser.c b/ext/nokogiri/html4_sax_push_parser.c
index 6955c0dcdb..cab3d31756 100644
--- a/ext/nokogiri/html4_sax_push_parser.c
+++ b/ext/nokogiri/html4_sax_push_parser.c
@@ -30,7 +30,7 @@ native_write(VALUE self, VALUE _chunk, VALUE _last_chunk)
 
   Nokogiri_structured_error_func_restore(&handler_state);
 
-  if ((status != 0) && !(ctx->options & XML_PARSE_RECOVER)) {
+  if ((status != 0) && !(xmlCtxtGetOptions(ctx) & XML_PARSE_RECOVER)) {
     // TODO: there appear to be no tests for this block
     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;
 
-  sax = noko_sax_handler_unwrap(_xml_sax);
+  sax = noko_xml_sax_parser_unwrap(_xml_sax);
 
   if (_filename != Qnil) { filename = StringValueCStr(_filename); }
 
@@ -79,7 +79,6 @@ initialize_native(VALUE self, VALUE _xml_sax, VALUE _filename,
 
   ctx->userData = NOKOGIRI_SAX_TUPLE_NEW(ctx, self);
 
-  ctx->sax2 = 1;
   DATA_PTR(self) = ctx;
   return self;
 }
diff --git a/ext/nokogiri/libxml2_polyfill.c b/ext/nokogiri/libxml2_polyfill.c
new file mode 100644
index 0000000000..70ed189e7c
--- /dev/null
+++ b/ext/nokogiri/libxml2_polyfill.c
@@ -0,0 +1,97 @@
+#include <nokogiri.h>
+
+#ifndef HAVE_XMLCTXTSETOPTIONS
+/* based on libxml2-2.14.0-dev (1d8bd126) parser.c xmlCtxtSetInternalOptions */
+int
+xmlCtxtSetOptions(xmlParserCtxtPtr ctxt, int options)
+{
+  int keepMask = 0;
+  int allMask;
+
+  if (ctxt == NULL) {
+    return (-1);
+  }
+
+  /*
+   * XInclude options aren't handled by the parser.
+   *
+   * XML_PARSE_XINCLUDE
+   * XML_PARSE_NOXINCNODE
+   * XML_PARSE_NOBASEFIX
+   */
+  allMask = XML_PARSE_RECOVER |
+            XML_PARSE_NOENT |
+            XML_PARSE_DTDLOAD |
+            XML_PARSE_DTDATTR |
+            XML_PARSE_DTDVALID |
+            XML_PARSE_NOERROR |
+            XML_PARSE_NOWARNING |
+            XML_PARSE_PEDANTIC |
+            XML_PARSE_NOBLANKS |
+#ifdef LIBXML_SAX1_ENABLED
+            XML_PARSE_SAX1 |
+#endif
+            XML_PARSE_NONET |
+            XML_PARSE_NODICT |
+            XML_PARSE_NSCLEAN |
+            XML_PARSE_NOCDATA |
+            XML_PARSE_COMPACT |
+            XML_PARSE_OLD10 |
+            XML_PARSE_HUGE |
+            XML_PARSE_OLDSAX |
+            XML_PARSE_IGNORE_ENC |
+            XML_PARSE_BIG_LINES;
+
+  ctxt->options = (ctxt->options & keepMask) | (options & allMask);
+
+  /*
+   * For some options, struct members are historically the source
+   * of truth. The values are initalized from global variables and
+   * old code could also modify them directly. Several older API
+   * functions that don't take an options argument rely on these
+   * deprecated mechanisms.
+   *
+   * Once public access to struct members and the globals are
+   * disabled, we can use the options bitmask as source of
+   * truth, making all these struct members obsolete.
+   *
+   * The XML_DETECT_IDS flags is misnamed. It simply enables
+   * loading of the external subset.
+   */
+  ctxt->recovery = (options & XML_PARSE_RECOVER) ? 1 : 0;
+  ctxt->replaceEntities = (options & XML_PARSE_NOENT) ? 1 : 0;
+  ctxt->loadsubset = (options & XML_PARSE_DTDLOAD) ? XML_DETECT_IDS : 0;
+  ctxt->loadsubset |= (options & XML_PARSE_DTDATTR) ? XML_COMPLETE_ATTRS : 0;
+  ctxt->validate = (options & XML_PARSE_DTDVALID) ? 1 : 0;
+  ctxt->pedantic = (options & XML_PARSE_PEDANTIC) ? 1 : 0;
+  ctxt->keepBlanks = (options & XML_PARSE_NOBLANKS) ? 0 : 1;
+  ctxt->dictNames = (options & XML_PARSE_NODICT) ? 0 : 1;
+
+  /*
+   * Changing SAX callbacks is a bad idea. This should be fixed.
+   */
+  if (options & XML_PARSE_NOBLANKS) {
+    ctxt->sax->ignorableWhitespace = xmlSAX2IgnorableWhitespace;
+  }
+  if (options & XML_PARSE_NOCDATA) {
+    ctxt->sax->cdataBlock = NULL;
+  }
+  if (options & XML_PARSE_HUGE) {
+    if (ctxt->dict != NULL) {
+      xmlDictSetLimit(ctxt->dict, 0);
+    }
+  }
+
+  ctxt->linenumbers = 1;
+
+  return (options & ~allMask);
+}
+#endif
+
+#ifndef HAVE_XMLCTXTGETOPTIONS
+int
+xmlCtxtGetOptions(xmlParserCtxtPtr ctxt)
+{
+  return (ctxt->options);
+}
+#endif
diff --git a/ext/nokogiri/nokogiri.c b/ext/nokogiri/nokogiri.c
index ea078f0964..d97413e5c2 100644
--- a/ext/nokogiri/nokogiri.c
+++ b/ext/nokogiri/nokogiri.c
@@ -46,6 +46,7 @@ void noko_init_html_element_description(void);
 void noko_init_html_entity_lookup(void);
 void noko_init_html_sax_parser_context(void);
 void noko_init_html_sax_push_parser(void);
+void noko_init_html4_sax_parser(void);
 void noko_init_gumbo(void);
 void noko_init_test_global_handlers(void);
 
@@ -244,7 +245,10 @@ Init_nokogiri(void)
   noko_init_xml_namespace();
   noko_init_xml_node_set();
   noko_init_xml_reader();
+
   noko_init_xml_sax_parser();
+  noko_init_html4_sax_parser();
+
   noko_init_xml_xpath_context();
   noko_init_xslt_stylesheet();
   noko_init_html_element_description();
diff --git a/ext/nokogiri/nokogiri.h b/ext/nokogiri/nokogiri.h
index 7a2e070af8..95e3ab366f 100644
--- a/ext/nokogiri/nokogiri.h
+++ b/ext/nokogiri/nokogiri.h
@@ -56,6 +56,14 @@
 
 #include <libexslt/exslt.h>
 
+/* libxml2_polyfill.c */
+#ifndef HAVE_XMLCTXTSETOPTIONS
+int xmlCtxtSetOptions(xmlParserCtxtPtr ctxt, int options);
+#endif
+#ifndef HAVE_XMLCTXTGETOPTIONS
+int xmlCtxtGetOptions(xmlParserCtxtPtr ctxt);
+#endif
+
 #define XMLNS_PREFIX "xmlns"
 #define XMLNS_PREFIX_LEN 6 /* including either colon or \0 */
 
@@ -141,6 +149,7 @@ NOKOPUBVAR VALUE cNokogiriXsltStylesheet ;
 NOKOPUBVAR VALUE cNokogiriHtml4Document ;
 NOKOPUBVAR VALUE cNokogiriHtml4SaxPushParser ;
 NOKOPUBVAR VALUE cNokogiriHtml4ElementDescription ;
+NOKOPUBVAR VALUE cNokogiriHtml4SaxParser;
 NOKOPUBVAR VALUE cNokogiriHtml4SaxParserContext;
 NOKOPUBVAR VALUE cNokogiriHtml5Document ;
 
@@ -196,7 +205,7 @@ xmlDocPtr noko_xml_document_unwrap(VALUE rb_document);
 NOKOPUBFUN VALUE Nokogiri_wrap_xml_document(VALUE klass,
     xmlDocPtr doc); /* deprecated. use noko_xml_document_wrap() instead. */
 
-xmlSAXHandlerPtr noko_sax_handler_unwrap(VALUE rb_sax_handler);
+xmlSAXHandlerPtr noko_xml_sax_parser_unwrap(VALUE rb_sax_handler);
 
 xmlParserCtxtPtr noko_xml_sax_push_parser_unwrap(VALUE rb_parser);
 
diff --git a/ext/nokogiri/xml_node.c b/ext/nokogiri/xml_node.c
index 0c59ae49e4..5c5e4e9c19 100644
--- a/ext/nokogiri/xml_node.c
+++ b/ext/nokogiri/xml_node.c
@@ -2081,7 +2081,7 @@ dump_html(VALUE self)
 
   buf = xmlBufferCreate() ;
   htmlNodeDump(buf, node->doc, node);
-  html = NOKOGIRI_STR_NEW2(buf->content);
+  html = NOKOGIRI_STR_NEW2(xmlBufferContent(buf));
   xmlBufferFree(buf);
   return html ;
 }
diff --git a/ext/nokogiri/xml_sax_parser.c b/ext/nokogiri/xml_sax_parser.c
index 989ad9eb30..5e84be5661 100644
--- a/ext/nokogiri/xml_sax_parser.c
+++ b/ext/nokogiri/xml_sax_parser.c
@@ -16,30 +16,28 @@ start_document(void *ctx)
 
   xmlParserCtxtPtr ctxt = NOKOGIRI_SAX_CTXT(ctx);
 
-  if (NULL != ctxt && ctxt->html != 1) {
-    if (ctxt->standalone != -1) { /* -1 means there was no declaration */
-      VALUE encoding = Qnil ;
-      VALUE standalone = Qnil;
-      VALUE version;
-      if (ctxt->encoding) {
-        encoding = NOKOGIRI_STR_NEW2(ctxt->encoding) ;
-      } else if (ctxt->input && ctxt->input->encoding) {
-        encoding = NOKOGIRI_STR_NEW2(ctxt->input->encoding) ;
-      }
-
-      version = ctxt->version ? NOKOGIRI_STR_NEW2(ctxt->version) : Qnil;
-
-      switch (ctxt->standalone) {
-        case 0:
-          standalone = NOKOGIRI_STR_NEW2("no");
-          break;
-        case 1:
-          standalone = NOKOGIRI_STR_NEW2("yes");
-          break;
-      }
-
-      rb_funcall(doc, id_xmldecl, 3, version, encoding, standalone);
+  if (ctxt->standalone != -1) { /* -1 means there was no declaration */
+    VALUE encoding = Qnil ;
+    VALUE standalone = Qnil;
+    VALUE version;
+    if (ctxt->encoding) {
+      encoding = NOKOGIRI_STR_NEW2(ctxt->encoding) ;
+    } else if (ctxt->input && ctxt->input->encoding) {
+      encoding = NOKOGIRI_STR_NEW2(ctxt->input->encoding) ;
     }
+
+    version = ctxt->version ? NOKOGIRI_STR_NEW2(ctxt->version) : Qnil;
+
+    switch (ctxt->standalone) {
+      case 0:
+        standalone = NOKOGIRI_STR_NEW2("no");
+        break;
+      case 1:
+        standalone = NOKOGIRI_STR_NEW2("yes");
+        break;
+    }
+
+    rb_funcall(doc, id_xmldecl, 3, version, encoding, standalone);
   }
 
   rb_funcall(doc, id_start_document, 0);
@@ -282,10 +280,9 @@ static const rb_data_type_t xml_sax_handler_type = {
 };
 
 static VALUE
-allocate(VALUE klass)
+noko_xml_sax_parser_initialize(VALUE self)
 {
-  xmlSAXHandlerPtr handler;
-  VALUE self = TypedData_Make_Struct(klass, xmlSAXHandler, &xml_sax_handler_type, handler);
+  xmlSAXHandlerPtr handler = noko_xml_sax_parser_unwrap(self);
 
   handler->startDocument = start_document;
   handler->endDocument = end_document;
@@ -304,8 +301,15 @@ allocate(VALUE klass)
   return self;
 }
 
+static VALUE
+noko_xml_sax_parser_allocate(VALUE klass)
+{
+  xmlSAXHandlerPtr handler;
+  return TypedData_Make_Struct(klass, xmlSAXHandler, &xml_sax_handler_type, handler);
+}
+
 xmlSAXHandlerPtr
-noko_sax_handler_unwrap(VALUE rb_sax_handler)
+noko_xml_sax_parser_unwrap(VALUE rb_sax_handler)
 {
   xmlSAXHandlerPtr c_sax_handler;
   TypedData_Get_Struct(rb_sax_handler, xmlSAXHandler, &xml_sax_handler_type, c_sax_handler);
@@ -317,7 +321,9 @@ noko_init_xml_sax_parser(void)
 {
   cNokogiriXmlSaxParser = rb_define_class_under(mNokogiriXmlSax, "Parser", rb_cObject);
 
-  rb_define_alloc_func(cNokogiriXmlSaxParser, allocate);
+  rb_define_alloc_func(cNokogiriXmlSaxParser, noko_xml_sax_parser_allocate);
+
+  rb_define_private_method(cNokogiriXmlSaxParser, "initialize_native", noko_xml_sax_parser_initialize, 0);
 
   id_start_document = rb_intern("start_document");
   id_end_document = rb_intern("end_document");
diff --git a/ext/nokogiri/xml_sax_parser_context.c b/ext/nokogiri/xml_sax_parser_context.c
index 87fd9f50d9..61468c48cb 100644
--- a/ext/nokogiri/xml_sax_parser_context.c
+++ b/ext/nokogiri/xml_sax_parser_context.c
@@ -155,7 +155,7 @@ parse_with(VALUE self, VALUE sax_handler)
   }
 
   ctxt = noko_xml_sax_parser_context_unwrap(self);
-  sax = noko_sax_handler_unwrap(sax_handler);
+  sax = noko_xml_sax_parser_unwrap(sax_handler);
 
   ctxt->sax = sax;
   ctxt->userData = (void *)NOKOGIRI_SAX_TUPLE_NEW(ctxt, sax_handler);
@@ -177,12 +177,17 @@ parse_with(VALUE self, VALUE sax_handler)
 static VALUE
 set_replace_entities(VALUE self, VALUE value)
 {
+  int error;
   xmlParserCtxtPtr ctxt = noko_xml_sax_parser_context_unwrap(self);
 
-  if (Qfalse == value) {
-    ctxt->replaceEntities = 0;
+  if (RB_TEST(value)) {
+    error = xmlCtxtSetOptions(ctxt, xmlCtxtGetOptions(ctxt) | XML_PARSE_NOENT);
   } else {
-    ctxt->replaceEntities = 1;
+    error = xmlCtxtSetOptions(ctxt, xmlCtxtGetOptions(ctxt) & ~XML_PARSE_NOENT);
+  }
+
+  if (error) {
+    rb_raise(rb_eRuntimeError, "failed to set parser context options (%x)", error);
   }
 
   return value;
@@ -200,10 +205,10 @@ get_replace_entities(VALUE self)
 {
   xmlParserCtxtPtr ctxt = noko_xml_sax_parser_context_unwrap(self);
 
-  if (0 == ctxt->replaceEntities) {
-    return Qfalse;
-  } else {
+  if (xmlCtxtGetOptions(ctxt) & XML_PARSE_NOENT) {
     return Qtrue;
+  } else {
+    return Qfalse;
   }
 }
 
@@ -255,12 +260,17 @@ column(VALUE self)
 static VALUE
 set_recovery(VALUE self, VALUE value)
 {
+  int error;
   xmlParserCtxtPtr ctxt = noko_xml_sax_parser_context_unwrap(self);
 
-  if (value == Qfalse) {
-    ctxt->recovery = 0;
+  if (RB_TEST(value)) {
+    error = xmlCtxtSetOptions(ctxt, xmlCtxtGetOptions(ctxt) | XML_PARSE_RECOVER);
   } else {
-    ctxt->recovery = 1;
+    error = xmlCtxtSetOptions(ctxt, xmlCtxtGetOptions(ctxt) & ~XML_PARSE_RECOVER);
+  }
+
+  if (error) {
+    rb_raise(rb_eRuntimeError, "failed to set parser context options (%x)", error);
   }
 
   return value;
@@ -278,10 +288,10 @@ get_recovery(VALUE self)
 {
   xmlParserCtxtPtr ctxt = noko_xml_sax_parser_context_unwrap(self);
 
-  if (ctxt->recovery == 0) {
-    return Qfalse;
-  } else {
+  if (xmlCtxtGetOptions(ctxt) & XML_PARSE_RECOVER) {
     return Qtrue;
+  } else {
+    return Qfalse;
   }
 }
 
diff --git a/ext/nokogiri/xml_sax_push_parser.c b/ext/nokogiri/xml_sax_push_parser.c
index b23f278791..dd131094e5 100644
--- a/ext/nokogiri/xml_sax_push_parser.c
+++ b/ext/nokogiri/xml_sax_push_parser.c
@@ -58,7 +58,7 @@ native_write(VALUE self, VALUE _chunk, VALUE _last_chunk)
   xmlSetStructuredErrorFunc(NULL, NULL);
 
   if (xmlParseChunk(ctx, chunk, size, Qtrue == _last_chunk ? 1 : 0)) {
-    if (!(ctx->options & XML_PARSE_RECOVER)) {
+    if (!(xmlCtxtGetOptions(ctx) & XML_PARSE_RECOVER)) {
       xmlErrorConstPtr e = xmlCtxtGetLastError(ctx);
       Nokogiri_error_raise(NULL, e);
     }
@@ -80,7 +80,7 @@ initialize_native(VALUE self, VALUE _xml_sax, VALUE _filename)
   const char *filename = NULL;
   xmlParserCtxtPtr ctx;
 
-  sax = noko_sax_handler_unwrap(_xml_sax);
+  sax = noko_xml_sax_parser_unwrap(_xml_sax);
 
   if (_filename != Qnil) { filename = StringValueCStr(_filename); }
 
@@ -97,7 +97,6 @@ initialize_native(VALUE self, VALUE _xml_sax, VALUE _filename)
 
   ctx->userData = NOKOGIRI_SAX_TUPLE_NEW(ctx, self);
 
-  ctx->sax2 = 1;
   DATA_PTR(self) = ctx;
   return self;
 }
@@ -109,18 +108,20 @@ get_options(VALUE self)
 
   ctx = noko_xml_sax_push_parser_unwrap(self);
 
-  return INT2NUM(ctx->options);
+  return INT2NUM(xmlCtxtGetOptions(ctx));
 }
 
 static VALUE
 set_options(VALUE self, VALUE options)
 {
+  int error;
   xmlParserCtxtPtr ctx;
 
   ctx = noko_xml_sax_push_parser_unwrap(self);
 
-  if (xmlCtxtUseOptions(ctx, (int)NUM2INT(options)) != 0) {
-    rb_raise(rb_eRuntimeError, "Cannot set XML parser context options");
+  error = xmlCtxtSetOptions(ctx, (int)NUM2INT(options));
+  if (error) {
+    rb_raise(rb_eRuntimeError, "Cannot set XML parser context options (%x)", error);
   }
 
   return Qnil;
@@ -136,14 +137,12 @@ set_options(VALUE self, VALUE options)
 static VALUE
 get_replace_entities(VALUE self)
 {
-  xmlParserCtxtPtr ctx;
-
-  ctx = noko_xml_sax_push_parser_unwrap(self);
+  xmlParserCtxtPtr ctxt = noko_xml_sax_push_parser_unwrap(self);
 
-  if (0 == ctx->replaceEntities) {
-    return Qfalse;
-  } else {
+  if (xmlCtxtGetOptions(ctxt) & XML_PARSE_NOENT) {
     return Qtrue;
+  } else {
+    return Qfalse;
   }
 }
 
@@ -157,14 +156,17 @@ get_replace_entities(VALUE self)
 static VALUE
 set_replace_entities(VALUE self, VALUE value)
 {
-  xmlParserCtxtPtr ctx;
-
-  ctx = noko_xml_sax_push_parser_unwrap(self);
+  int error;
+  xmlParserCtxtPtr ctxt = noko_xml_sax_push_parser_unwrap(self);
 
-  if (Qfalse == value) {
-    ctx->replaceEntities = 0;
+  if (RB_TEST(value)) {
+    error = xmlCtxtSetOptions(ctxt, xmlCtxtGetOptions(ctxt) | XML_PARSE_NOENT);
   } else {
-    ctx->replaceEntities = 1;
+    error = xmlCtxtSetOptions(ctxt, xmlCtxtGetOptions(ctxt) & ~XML_PARSE_NOENT);
+  }
+
+  if (error) {
+    rb_raise(rb_eRuntimeError, "failed to set parser context options (%x)", error);
   }
 
   return value;
diff --git a/lib/nokogiri/xml/sax/parser.rb b/lib/nokogiri/xml/sax/parser.rb
index 289aabe318..c9447f6496 100644
--- a/lib/nokogiri/xml/sax/parser.rb
+++ b/lib/nokogiri/xml/sax/parser.rb
@@ -73,6 +73,8 @@ def initialize(doc = Nokogiri::XML::SAX::Document.new, encoding = "UTF-8")
           @encoding = check_encoding(encoding)
           @document = doc
           @warned   = false
+
+          initialize_native unless Nokogiri.jruby?
         end
 
         ###
diff --git a/nokogiri.gemspec b/nokogiri.gemspec
index f4f3941b4e..faa875ecaa 100644
--- a/nokogiri.gemspec
+++ b/nokogiri.gemspec
@@ -150,8 +150,10 @@ Gem::Specification.new do |spec|
     "ext/nokogiri/html4_document.c",
     "ext/nokogiri/html4_element_description.c",
     "ext/nokogiri/html4_entity_lookup.c",
+    "ext/nokogiri/html4_sax_parser.c",
     "ext/nokogiri/html4_sax_parser_context.c",
     "ext/nokogiri/html4_sax_push_parser.c",
+    "ext/nokogiri/libxml2_polyfill.c",
     "ext/nokogiri/nokogiri.c",
     "ext/nokogiri/nokogiri.h",
     "ext/nokogiri/xml_attr.c",
diff --git a/test/xml/sax/test_parser.rb b/test/xml/sax/test_parser.rb
index 8551a3469c..832ab05e2e 100644
--- a/test/xml/sax/test_parser.rb
+++ b/test/xml/sax/test_parser.rb
@@ -49,15 +49,15 @@ class TestCase
           assert_nil(parser.document.xmldecls)
         end
 
-        it :test_xml_decl do
-          [
-            ['<?xml version="1.0" ?>', ["1.0"]],
-            ['<?xml version="1.0" encoding="UTF-8" ?>', ["1.0", "UTF-8"]],
-            ['<?xml version="1.0" standalone="yes"?>', ["1.0", "yes"]],
-            ['<?xml version="1.0" standalone="no"?>', ["1.0", "no"]],
-            ['<?xml version="1.0" encoding="UTF-8" standalone="no"?>', ["1.0", "UTF-8", "no"]],
-            ['<?xml version="1.0" encoding="ISO-8859-1" standalone="yes"?>', ["1.0", "ISO-8859-1", "yes"]],
-          ].each do |decl, value|
+        [
+          ['<?xml version="1.0" ?>', ["1.0"]],
+          ['<?xml version="1.0" encoding="UTF-8" ?>', ["1.0", "UTF-8"]],
+          ['<?xml version="1.0" standalone="yes"?>', ["1.0", "yes"]],
+          ['<?xml version="1.0" standalone="no"?>', ["1.0", "no"]],
+          ['<?xml version="1.0" encoding="UTF-8" standalone="no"?>', ["1.0", "UTF-8", "no"]],
+          ['<?xml version="1.0" encoding="ISO-8859-1" standalone="yes"?>', ["1.0", "ISO-8859-1", "yes"]],
+        ].each do |decl, value|
+          it "parses xml decl '#{decl}'" do
             parser = Nokogiri::XML::SAX::Parser.new(Doc.new)
 
             xml = "#{decl}\n<root />"