Re: Getting bison to ignore tokens... sometimes!

[Evan Lavelle]

If your parser knows that it's dealing with a statement, you could try 
lexical feedback (look up "lexical tie-ins" in the bison docs, and 
google "scanner feedback" or "lexical feedback"). The parser sets a flag 
to tell the lexer that it's inside a statement, and the lexer then 
returns the NL when the flag is set, and discards it otherwise.

I've never had much luck with this, because the lookahead can make it 
difficult - the lexer might already have decided what to do before the 
parser tells it. Still, might work for you.

-Evan