From: Park Heesob <phasis@...>
Date: 2003-04-18T15:00:58+09:00
Subject: Re: Roundoff problem with Float and Marshal

Hi,
----- Original Message -----
From: <nobu.nokada@softhome.net>
To: "ruby-talk ML" <ruby-talk@ruby-lang.org>
Sent: Friday, April 18, 2003 2:32 PM
Subject: Re: Roundoff problem with Float and Marshal


> Hi,
>
> At Fri, 18 Apr 2003 12:20:01 +0900,
> Yukihiro Matsumoto wrote:
> > It's the limitation of base 10 representation of float values that
> > Marshal uses.  We need to use binary representation like XDR to solve
> > this.  I'm not sure if it is worth for the cost of its implementation
> > and portability.  Float tend to be inexact, after all.
>
> I'd tried C99 style hexadecimal float.  I expect this is
> portable (except for depending on frexp() and ldexp()), but
> could not be loaded by current version.
>
> But still, strtod() problem remains.
>
>
How about pack unpack method? Not Portable?

For Float f,
  f == [f].pack("G").unpack("G")[0]
is always true.

Park Heesob