peep::a is not harmful for speed, but it does much les work then b ususally.

Both need to be replaced, I think.

Yang, the "replace subexpression with 0 or 1" pass can't really be done by a lexer (much less a regex). Can you add that to your clang_delta TODO list?

Of course, pointer-typed subexpressions should only by replaced with 0.

John

Reply via email to