Yes, it's architecturally dependent in a minor way.  In particular, 
it would work for J64 for integers.  For example, the following
would fail in J64:
   (o.1 0 2) +/ @: *  (o.3 2 $ 3 1 2 1 1 0)

The bug in x +/@:* y occurs when:
- x and y have a different number of elements.
- the argument with a smaller number of elements has a 0.
Unless: x and y are integers and the system is J64.
(In that case it skips the special and erroneous code.)

This has now been fixed.



----- Original Message -----
From: Tikkanz <[email protected]>
Date: Monday, January 10, 2011 11:26
Subject: Re: [Jbeta] Nag nag nag nag
To: Beta forum <[email protected]>

> Seems to work OK for me. Maybe it is architecture/processor dependent?
> 
>    JVERSION
> Engine: j701/2011-01-07/14:33
> Library: 7.01.030
> Platform: Linux 64
> Installer: j701a_linux64.sh
> InstallPath: /usr/local/lib/j64_701
> 
>    1 0 2 +/ @: *  (3 2 $ 3 1 2 1 1 0)
> 5 1
> 
> 
> On Tue, Jan 11, 2011 at 6:10 AM, June Kim (김창준) 
> <[email protected]> wrote:
> > Uh-oh. I've been doing financial calculations with J and the 
> following (old)
> > bug report is an alarm for me. Have to look at my codes again.
> >
> > What is the minimal condition for the bug?
> >
> > On Mon, Jan 10, 2011 at 7:02 PM, Roger Stokes <
> > [email protected]> wrote:
> >
> >> Still got this old bug in j701:
> >>
> >>   JVERSION
> >> Engine: j701/2011-01-07/14:33
> >> Library: 7.01.030
> >> Platform: Win 32
> >> Installer: j701a_win.exe
> >> InstallPath: c:/users/homer/j701
> >>
> >>         1 0 2 +/ @: *  (3 2 $ 3 1 2 1 1 0)
> >> 7 3
> >>
> >>         NB. The result 7 3 is wrong.
> >>         NB. The correct result is 5 1, shown by replacing +
> >>         NB. with an equivalent verb, say  +"0 .
> >>
> >>         1 0 2 +"0 / @: * (3 2 $ 3 1 2 1 1 0)
> >> 5 1
> >>
> >>         NB. This has been mentioned before, firstly
> >>         NB. by Patrick Van Beek on 14 6 2008.
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to