Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20030422194057.GA15164@openwall.com>
Date: Tue, 22 Apr 2003 23:40:57 +0400
From: Solar Designer <solar@...nwall.com>
To: owl-users@...ts.openwall.com
Subject: Re: Next Release

On Tue, Apr 22, 2003 at 11:29:48AM -0500, Steve Bremer wrote:
> Although I wouldn't mind seeing the 2.4 kernel become the default, 
> I've learned to really like the slow/steady/safe development pace 
> taken by Owl.

It's nice to know that.

> Heck, even current is extremely stable in my experience.

It is, but I don't view this as an obviously good thing.  When we
released 1.0, the plan was to start more aggressive development in
current and maintain a 1.0-stable for the security fixes.

In reality, we simply didn't have enough time to start doing all that
was planned for post-1.0 Owl right away, so current continued to
evolve at about the same pace that it had shortly before 1.0.  And it
became in all ways better than 1.0, -- including being even more
stable (well, we break things in our private cvs for a few days once
in a while, but I simply delay anoncvs/ftp updates in those cases).

So the new plan is to continue with safe changes only for a while
longer, make a 1.1 release, and only then hopefully do the major
updates that were originally planned for shortly after 1.0.  (I am
talking primarily of the gcc and glibc updates.)

To do these heavy changes now, without making another release first,
would be counter-productive because fewer people would then be able to
benefit from the stability improvements achieved in current so far.

> I do think 2.4 has some nice firewalling/networking features that 2.2 
> lacks, but I can live with 2.2 for awhile if it means Owl continues to 
> stay as stable and reliable as it has been.

To me, the primary reasons for Owl to move to 2.4.x aren't the more
extensive firewalling features, but rather availability of drivers
for newer hardware (most importantly disk controllers) and improved
kernel interfaces which allow to do neat things such as Olaf Kirch's
non-SUID/SGID traceroute that behaves more like the traditional one.

Thank you for your feedback, -- it really helps.

-- 
/sd

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.