Re: [Gcl-devel] Bug possibly related to integer type propagation?

[Camm Maguire]

Greetings!

Thanks!  Should be fixed now.  Not a type propagation problem, but
rather in the C implementation of fixnum_add.

Take care,

"Paul F. Dietz" <address@hidden> writes:

> Test MISC.461 failed
> Form: (FUNCALL (COMPILE NIL
>                          '(LAMBDA (A)
>                             (DECLARE (TYPE (INTEGER -26657952320 0) A))
>                             (DECLARE (OPTIMIZE (COMPILATION-SPEED 0)
>                                       (SPACE 3) (SPEED 3) (SAFETY 0)
>                                       (DEBUG 2)))
>                             (- A (ASH -1 (MIN 31 (- A))) -26715477)))
>                 -26179151369)
> Expected value: -24004952244
> Actual value: -28299919540.
> 
> 
> It's off by 2^32.
> 
>       Paul
> 
> 
> _______________________________________________
> Gcl-devel mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/gcl-devel
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah