Re: Crays

On Thu, Jul 26, 2001 at 11:49:12PM +0100, Nicholas Clark wrote:
> Sorry about the delays recently - I've been somewhat busy recently due
> to a variety of reasons.
> My inbox says that following patch makes op/numconvert pass tests
> 209,210,217,218 on Crays
> [and presumably anything else where 1.0 + (double) ~0 is less than ~0]
> 
> --- sv.c.orig	Thu Jul  5 05:44:59 2001
> +++ sv.c	Sat Jul  7 21:34:10 2001
> @@ -5829,7 +5829,7 @@
>        oops_its_int:
>  	if (SvIsUV(sv)) {
>  	    if (SvUVX(sv) == UV_MAX)
> -		sv_setnv(sv, (NV)UV_MAX + 1.0);
> +		sv_setnv(sv, UV_MAX_P1);

Thanks, applied.

>  	    else
>  		(void)SvIOK_only_UV(sv);
>  		++SvUVX(sv);
> 
> 
> I may have more time on my hands in the short term, so hopefully my do
> list will see some progress.
> [remaining numeric problems on Crays {any other platforms?}
>  64 bit pp_divide, pp_modulo, grokking 64 bit hex, octal and binary,
>  filters on coderef @INC, fixing ex::lib::zip because N I-S is a moving
>  target [about 150 miles], h2xs bugs and patches passed to me
>  and probably some things that slip my mind.]
> 
> Nicholas Clark
> -- 
> EAGAIN: http://plum.flirble.org/~nick/CV.html

-- 
$jhi++; # http://www.iki.fi/jhi/
        # There is this special biologist word we use for 'stable'.
        # It is 'dead'. -- Jack Cohen

• Crays by Nicholas Clark
• Re: Crays by Jarkko Hietaniemi