iteration variable type for a single float not derived

[Nikodemus Siivola]

  (defun test2 ()
    (let ((f most-positive-single-float))
      (tagbody
       :next
         (unless (> f 0.0)
           (go :end))
         (foo f)
         (setq f (/ f 2.0))
         (go :next)
     :end)))

uses GENERIC->, because the compiler doesn't realize F cannot be a double-float.

[Paul F. Dietz (paul-f-dietz)]

This no longer appears to be happening (1.4.11+).

[Douglas Katzman (dougk)]

this problem is still present but is not of medium importance