Float printing bug still around?

Or have I found something else?

I'm running 0.13.6 under OS X, and the following stupid code gives 
screwy printouts (but proper underlying numbers) for anything more than
about 15 iterations when using integer values for all the inputs:

(defun loanlength (amount interest payment)
   (let ((monthly-int (/ interest 1200)))
(do ((currentmonth 0 (1+ currentmonth))
      (balance amount (- balance (- payment (* balance monthly-int)))))
     ((< balance 0) currentmonth)
   (format t "~&Month ~D Balance ~,2F Equity Change ~,2F" currentmonth 	
           balance)
	  (- payment (* balance monthly-int))))))

The rational representing the balance for a typical example is something 
like

29479420866365225234014731893929685749/6058287395472553279488000000000000

when things go bad.

For the moment, "don't use integer arguments" seems like a perfectly 
fine fix, but I was under the impression from reading the archive that 
that bug had been stomped. Hence the question

thanks

paul