[Bug-openmcl] Is it a bug?

[Gary Byers]

Sorry; I just saw this a minute ago.  (I tend to file bug-openmcl mail
in with all of the messages that I get about spammmers trying to send
mail to bug-openmcl, which isn't good ...)

It's a bug in the PPC backend; it doesn't realize that a register's
getting clobbered.  I'll try to fix it in a minute or two.

(In the meantime, you can get avoid the bug by doing:

(defun make-graph (ld lst)
   (let ((len (length lst)))
     (let ((va (apply #'vector lst))
           (ga (make-array (list len len))))
       (dotimes (i len)
         (dotimes (j len)
           (DECLARE (NOTINLINE CCL::ASET)) ; <<< added
           (setf (aref ga i j) (funcall ld (svref va i) (svref va j)))))
       (values ga va))))

)

On Fri, 9 Nov 2007, Dmitri Hrapof wrote:

> Ralf Stoye wrote:
>> Tested without problem on
>> Version 1.1-r7313 (DarwinPPC32) and
>> Version 1.1-r7591 (DarwinPPC64)
> May it be a problem on LinuxPPC32?
> Is there any way for me to debug it or provide more information?
>
> It's really a pity I can't use OpenMCL as SBCL doesn't support threads
> on PPC...
>
> Sincerely yours,
> Dmitri
> _______________________________________________
> Bug-openmcl mailing list
> Bug-openmcl at clozure.com
> http://clozure.com/mailman/listinfo/bug-openmcl
>
>