From: Mathieu Bouchard <matju@...>
Date: 2002-02-03T05:21:02+09:00
Subject: Re: smallest non-zero number (and other limits)


On Sat, 2 Feb 2002, Yukihiro Matsumoto wrote:
>  (b) Ruby only has double precision float (Float class), why do we
>      need these FLT_* constants?

Remove the type prefix; keep only the ones for "double". the constants for
the other types will get defined by the same names in different classes
(?) _whenever_ they become relevant.

(otoh, NArray uses single floats too; don't know how that would fit.)

>  (c) I still think we should do something other than just ommiting on
>      non ANSI platforms if we do check.  Or just abandon non ANSI C
>      platforms.

It should be expected that all supported platforms are at least ANSI C,
and keep the compatibility hacks outside of the main source
code.

e.g.1. In this case you'd just assume <float.h> is there, and if not,
someone will provide that file with at least semi-sensible values in it.

e.g.2. I'd prefer the Ruby code to be in canonic ANSI C, and whoever
really needs Ruby on a machine not supporting GCC (nor any other ANSI C
compiler) should use a source translator to convert back to K&R C.

________________________________________________________________
Mathieu Bouchard                   http://hostname.2y.net/~matju