LibWeb: Escape HTML text fragments with multi-byte code point awareness

The UTF-8 encoding of U+00A0 (NBSP) is the bytes 0xc2 0xa0. By looping
over the string to escape byte-by-byte, we replace the second byte with
" ", but leave the first byte in the resulting text. This creates
an invalid UTF-8 string, with a lone leading byte.

Userland/Libraries/LibWeb/HTML/Parser/HTMLParser.cpp

@@ -3586,23 +3586,23 @@ DeprecatedString HTMLParser::serialize_html_fragment(DOM::Node const& node)
  auto escape_string = [](StringView string, AttributeMode attribute_mode) -> DeprecatedString {
  // https://html.spec.whatwg.org/multipage/parsing.html#escapingString
  StringBuilder builder;
- for (auto& ch : string) {
+ for (auto code_point : Utf8View { string }) {
  // 1. Replace any occurrence of the "&" character by the string "&".
- if (ch == '&')
+ if (code_point == '&')
  builder.append("&"sv);
  // 2. Replace any occurrences of the U+00A0 NO-BREAK SPACE character by the string " ".
- else if (ch == '\xA0')
+ else if (code_point == 0xA0)
  builder.append(" "sv);
  // 3. If the algorithm was invoked in the attribute mode, replace any occurrences of the """ character by the string """.
- else if (ch == '"' && attribute_mode == AttributeMode::Yes)
+ else if (code_point == '"' && attribute_mode == AttributeMode::Yes)
  builder.append("""sv);
  // 4. If the algorithm was not invoked in the attribute mode, replace any occurrences of the "<" character by the string "&lt;", and any occurrences of the ">" character by the string "&gt;".
- else if (ch == '<' && attribute_mode == AttributeMode::No)
+ else if (code_point == '<' && attribute_mode == AttributeMode::No)
  builder.append("&lt;"sv);
- else if (ch == '>' && attribute_mode == AttributeMode::No)
+ else if (code_point == '>' && attribute_mode == AttributeMode::No)
  builder.append("&gt;"sv);
  else
- builder.append(ch);
+ builder.append_code_point(code_point);
  }
  return builder.to_deprecated_string();
  };