Broken html when crossref matches `#<` without linking `<`

[4lllex]

https://docs.ruby-lang.org/en/4.0/Encoding/Converter.html#method-i-convpath

<p>In the above example, [#&lt;Encoding:ISO-8859-1&gt;,
 <encoding:utf-8&gt;] means="" a="" converter="" from="" iso-8859-1="" to="" utf-8.="" “crlf_newline”="" newline="" lf="" crlf.<="" p="">
 </encoding:utf-8&gt;]>
</p>

Second #<Encoding with a leading space is matched:

"In the above example, [#<Encoding:ISO-8859-1>, #<Encoding:UTF-8>] means"
  .match(RDoc::CrossReference::CROSSREF_REGEXP, 2)

# => #<MatchData " #<" 1:"#<" 2:nil 3:nil 4:nil 5:"<" 6:nil>

"<" doesn't get linked by handle_regexp_CROSSREF and assigns handled as unescaped text "<" which breaks html:

rdoc/lib/rdoc/markup/formatter.rb

Lines 113 to 115 in d6422f4

	if matched_name
	handled = public_send(:"handle_regexp_#{matched_name}", matched_text)
	output << [handled, true]