From: Micah Wylde <wyldeone@...>
Date: 2007-03-02T10:19:01+09:00
Subject: Re: Floats that are NaN have strange behavior

The expected behavior is present in ruby 1.8.4 (2005-12-24), as well.

$ ruby -ve 'p [(0.0/0.0).zero?, (0.0/0.0).nonzero?]'
ruby 1.8.4 (2005-12-24) [i686-darwin8.8.3]
[false, NaN]

On 3/1/07, Tadashi Saito <shiba@mail2.accsnet.ne.jp> wrote:
> Hi Jonas,
>
> On Fri, 2 Mar 2007 03:10:44 +0900
> Jonas Kongslund <jonas@kongslund.dk> wrote:
>
> > I'm playing around with Ruby and noticed the following behavior on the
> > float 0.0/0.0 which is NaN:
> >
> > (0.0/0.0).zero? evaluates to true. I would expect false.
> > (0.0/0.0).nonzero? evaluates to nil. I would expect NaN.
> >
> > Is it a bug?
>
> I guess so, btw, what's your ruby's version and environment?
> `ruby -v` will help.
>
> Here is my case:
>
> $ ./ruby-1.8.6/bin/ruby-1.8.6  -ve 'p [(0.0/0.0).zero?,
> (0.0/0.0).nonzero?]'
> ruby 1.8.6 (2007-02-28 patchlevel 0) [i686-linux]
> [false, NaN]
> $ ruby-trunk -ve 'p [(0.0/0.0).zero?, (0.0/0.0).nonzero?]'
> ruby 1.9.0 (2007-03-01 patchlevel 0) [i686-linux]
> [false, NaN]
>
> --
> Tadashi Saito
>
>


-- 
Micah Wylde