The problem: 

A full parser plus lexer consists of of a set of three different things:

  . Grammar rules for the LALR(1) parser
  . Regex macros (like 'digit')
  . Tokens

Later maybe even:
 
  . Macros for the grammar section

How to tell these apart?

-------------------------------------------------------------------------------
variant i -- magic

    (    sum -> product
             -> product "+" sum => (+ $1 $3))
    (product -> factor
             -> factor "*" factor => (* $1 $3))
    ( factor -> :number
             -> "(" sum ")" => $2)

    (:number -> (+ digit) => (parse-integer $$))
    (  digit -> (range #\0 #\9))

    (:ignore -> #\space)

-------------------------------------------------------------------------------
variant ii -- special 'token' rule and some magic

    (    sum -> product
             -> product "+" sum => (+ $1 $3))
    (product -> factor
             -> factor "*" factor => (* $1 $3))
    ( factor -> number
             -> "(" sum ")" => $2)

    (token   -> number
             -> space)

    (number -> (+ digit) => (parse-integer $$))
    ( digit -> (range #\0 #\9))

    (space  -> #\space)

-------------------------------------------------------------------------------
variant iii -- separate sections

:rules
    (    sum -> product
             -> product "+" sum => (+ $1 $3))
    (product -> factor
             -> factor "*" factor => (* $1 $3))
    ( factor -> :number
             -> "(" sum ")" => $2)

:defs
    (  digit -> (range #\0 #\9))

:tokens
    (:number -> (+ digit) => (parse-integer $$))
    (:ignore -> #\space)

-------------------------------------------------------------------------------
variant iv -- each 'rule' prefixed

    (:rule     sum -> product
                   -> product "+" sum => (+ $1 $3))
    (:rule product -> factor
                   -> factor "*" factor => (* $1 $3))
    (:rule  factor -> number
                   -> "(" sum ")" => $2)

    (:macro  digit -> (range #\0 #\9))

    (:token number -> (+ digit) => (parse-integer $$))
    (:token ignore -> #\space)

-------------------------------------------------------------------------------
variant v -- each 'rule' infixed

    (   sum --> product
            --> product "+" sum => (+ $1 $3))
    (roduct --> factor
            --> factor "*" factor => (* $1 $3))
    (factor --> number
            --> "(" sum ")" => $2)

    ( digit =  (range #\0 #\9))

    (number -> (+ digit) => (parse-integer $$))
    (ignore -> #\space)