Major debugging breakage

[law@redhat.com]

  > > Note that we call c_expand_start_cond (which records the line number for
  > > the statement) *after* we've parsed the entire expression.  Thus the
  > > recorded line number is wrong.
  > > 
  > > At least that's my theory :-)
  > 
  > It looks very likely to me.
Agreed -- particularly since I've managed to bastardize my tree so that
we build the base IF and WHILE nodes before we have finished parsing the
condition.  With that change the debug symbols look correct.

I've got to clean that code up and submit it (tomorrow AM since the tests
will be running through the night).

jeff