Keeping track of token locations in a parser

Folks,

I understand that you can build a lexer on top of Edi's cl-ppcre.  
There's in fact a macro somewhere that shows how to do this.

How would you keep track of token locations in a lexer or parser,  
though, for a later type-checking phase?

I'm parsing non-Lisp code.

	Thanks, Joel

--
http://wagerlabs.com/