From: Andrew Walrond <andrew@...>
Date: 2005-02-02T05:02:40+09:00
Subject: Re: [BUG] Segmentation fault

On Tuesday 01 February 2005 18:27, Yukihiro Matsumoto wrote:
> Hi,
>
> In message "Re: [BUG] Segmentation fault"
>
>     on Wed, 2 Feb 2005 02:03:05 +0900, Eric Hodel <drbrain@segment7.net> 
writes:
> |> Hope it might mean something to someone :)
> |
> |It would mean more with a backtrace, and if your Ruby was compiled with
> |debugging symbols...
>
> And if the error reproducing Ruby program (along with input data) is
> supplied.
>

Sorry guys. I always try my best to provide a patch, a testcase etc (google me 
and you'll see what I mean) But I'm afraid its another of those inherently 
unreproducible bugs. You're welcome to the code, but It's been run thousands 
of times with the same dataset andlib/modules/fglrx/build_mod/ this is the 
first (and last) time I've seen this particular segfault...

I am pushing ruby quite hard with Rubyx, and do seem to turn up these 
unreproducible bugs (search back a couple months and you'll find the last one 
I was shouting about which turned out to be a nasty garbage collector bug, 
and really had me pulling my hair out :))

The purpose of my brief post was because, from (...long...) experience, it 
often useful for the author of code in question to have a quick look at 
possible failure modes in that area. I guess thats why the segfault notice 
has file name and linenumber...

If I can turn anything else up, I'll be sure to post it here.

Andrew Walrond