[CM] 'drunk' behavior

[Rick Taube]

indeed. the bug is that the value is bounded (between (- width)  
width) which means that there is 1 more neg. number than positive :/
if you change that to (between (- width) (+ width 1) ) then it  
behaves correctly ie drunkvec averages to the initial value:

(defun drunkvec (startpit len &optional (stepsize 1))
    (loop repeat len
      for r = startpit then (drunk r stepsize)
      collect r))

(defun avr (l) (floor (apply #'+ l) (length l)))


? (avr (loop repeat 30 append (last (drunkvec 50 20))))
=> 49
? (avr (loop repeat 30 append (last (drunkvec 50 20))))
=> 51
? (avr (loop repeat 30 append (last (drunkvec 50 20))))
=> 50

this is fixed in cvs (rev 1.14 of data.scm). thanks drew!