Fix: Set encoding in XML.parse_html explicitly to UTF-8 (crystal-la…

…ng#13705)

spec/std/xml/xml_spec.cr

@@ -419,6 +419,38 @@ describe XML do
     assert_prints node.to_xml, %(<p>&lt;foo&gt;</p>)
   end
 
+  it "parses HTML UTF-8 from memory (#13703)" do
+    doc = XML.parse_html("<p>České psaní</p>")
+
+    node = doc.root.try(&.children.first).should_not be_nil
+
+    node.text.should eq "České psaní"
+  end
+
+  it "parses HTML UTF-8 from IO (#13703)" do
+    doc = XML.parse_html(IO::Memory.new("<p>České psaní</p>"))
+
+    node = doc.root.try(&.children.first).should_not be_nil
+
+    node.text.should eq "České psaní"
+  end
+
+  it "parses XML UTF-8 from memory (#13703)" do
+    doc = XML.parse("<p>České psaní</p>")
+
+    node = doc.root.try(&.children.first).should_not be_nil
+
+    node.text.should eq "České psaní"
+  end
+
+  it "parses XML UTF-8 from IO (#13703)" do
+    doc = XML.parse(IO::Memory.new("<p>České psaní</p>"))
+
+    node = doc.root.try(&.children.first).should_not be_nil
+
+    node.text.should eq "České psaní"
+  end
+
   it "gets empty content" do
     doc = XML.parse("<foo/>")
     doc.children.first.content.should eq("")

src/xml.cr

@@ -78,7 +78,7 @@ module XML
   # See `HTMLParserOptions.default` for default options.
   def self.parse_html(string : String, options : HTMLParserOptions = HTMLParserOptions.default) : Node
     raise XML::Error.new("Document is empty", 0) if string.empty?
-    from_ptr { LibXML.htmlReadMemory(string, string.bytesize, nil, nil, options) }
+    from_ptr { LibXML.htmlReadMemory(string, string.bytesize, nil, "utf-8", options) }
   end
 
   # Parses an HTML document from *io* with *options* into an `XML::Node`.
@@ -92,7 +92,7 @@ module XML
       ->(ctx) { 0 },
       Box(IO).box(io),
       nil,
-      nil,
+      "utf-8",
       options,
     ) }
   end