[Sbcl-devel] another type bug?

[John K. <jk2...@ya...>]

Hello,

In the 0.8.5 release, the following won't compile, though
it compiles fine under CMUCL.

Regards,
John


(defun find-lc-for-e  (phi-func r-max energy)
  (declare (type double-float r-max energy)
	   (type (function (double-float) double-float) phi-func))
  (loop with best-l2 of-type double-float = 0d0
	;; need to add phi0 to phi because we define phi (and E) to
	;; have a minimum value of 0
	with phi0 = (funcall phi-func 0d0)
	for r of-type double-float from 0d0 to r-max by (/ r-max 1d5)
	for l2 of-type double-float = (* 2d0 r r
					 (- energy
					    (+ phi0 (funcall phi-func r))))
	do
	(setf best-l2 (max l2 best-l2))
	finally (return (sqrt best-l2))))
	

The error is:
;;;;;;;;;;;;;;;
;     (LOOP WITH BEST-L2 = 0.0d0 WITH PHI0 = (FUNCALL PHI-FUNC 0.0d0) FOR ...)
; 
; caught error:
;   (in macroexpansion of (loop with best-l2 ...))
;   (hint: For more precise location, try *BREAK-ON-SIGNALS*.)
;   0 can't be converted to type (and double-float (real (0))).

;;;;;;;;;;;;;;;;





;;;;;;;;;;;;;;;;;;;;;;;
PS: in my last mail I think that some 'with' loop clauses
should have been 'for' clauses, but this doesn't change the fact
that the compiler got a 'shouldn't happen' type bug



__________________________________
Do you Yahoo!?
Protect your identity with Yahoo! Mail AddressGuard
http://antispam.yahoo.com/whatsnewfree