Re: [Gcl-devel] funcall bug (not ansi !! Yay)

[Camm Maguire]

Hi Peter!  I think you are right -- in fact its hard for me to imagine
you being wrong!  Plus it checks out with maxima and acl2.  So its
going in.  Thanks again!!!

take care,

Peter Wood <address@hidden> writes:

> --FL5UXtIhxfXey3p5
> Content-Type: text/plain; charset=us-ascii
> Content-Disposition: inline
> 
> Hi
> 
> Here's a refreshingly non-ansi bug.  It exists on CVS GCL (downloaded
> now), and 2.4.3.  Two examples:
> 
> >(funcall #'eval '(+ 1 1))
> 
> =>#<compiled-function EVAL>
> 
> >(mapcar #'eval '((+ 1 1) (funcall #'eval '(+ 1 1))))
> 
> =>(2 #<compiled-function EVAL>)
> 
> 
> Weird, huh?
> 
> I 'fixed' it with the appended patch. DISCLAIMER: the fix is so
> obvious there has to be something wrong with it too.  The line I
> changed goes all the way back to KCL!?
> 
> Does anyone have any notes or documentation for the internal types:
> t_afun, t_sfun, etc?  Or just knowledge they would share with the list?
> 
> Regards,
> Peter
> --FL5UXtIhxfXey3p5
> Content-Type: text/plain; charset=us-ascii
> Content-Disposition: attachment; filename="eval.patch"
> 
> --- o/eval.c~ Tue Sep 24 21:01:45 2002
> +++ o/eval.c  Thu Nov 21 17:28:51 2002
> @@ -1175,7 +1175,8 @@
>  
>       /* 1 args */
>       lex_new();
> -     eval(vs_base[0]);
> +     /*eval(vs_base[0]);*/
> +     eval(x0);
>       lex_env = lex;
>       return Ivs_values();
>  }
> 
> --FL5UXtIhxfXey3p5--
> 
> 
> _______________________________________________
> Gcl-devel mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/gcl-devel
> 
> 

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