Suggested resolution:

Change in 5.5 [lex.pptoken] bullet 5.4.1 as follows:

• ...
• Otherwise, the next preprocessing token is the longest sequence of characters that could constitute a preprocessing token, even if that would cause further lexical analysis to fail, except that
  • a string-literal or preprocessing-op-or-punc preprocessing token is never formed when a header-name preprocessing token can be formed, and
  • ...

(From submission #910.)

Consider:

  #include <=>

Is this lexed as a header-name or a punctuator?