From: nobu.nokada@...
Date: 2003-08-16T21:54:10+09:00
Subject: Re: _id2ref bug? (another break)

Hi,

At Sat, 16 Aug 2003 20:08:44 +0900,
ts wrote:
> n> Ryan is correct.
> 
>  not for me. There is a difference between protecting ruby againt coredump
>  and writing stupid code.

I agree that the protection should work.

> n>   $ ruby -e 'x = ""; eval("x"); y = ""; x = [x.id, y.id]; x.min.step(x.max, 10){|i| p ObjectSpace._id2ref(i)}'
> 
>  I'll never trust a module which have such code, and I hope that *no*
>  modules in the standard distribution do something like this.

I know nothing about his code, and this is pure ruby code.  It
means that the protection doesn't work enough correctly now.

-- 
Nobu Nakada