Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 20 Jun 2014 00:27:40 +0400
From: "(GalaxyMaster)" <galaxy@...nwall.com>
To: owl-dev@...ts.openwall.com
Subject: Re: massive Owl userland updates

All,

On Sat, Jun 14, 2014 at 03:43:07PM +0400, (GalaxyMaster) wrote:
> The first batch will be auto* tools and m4 to bring the build system to
> the level to be able to built other stuff.

I'm still alive :).  It turned out that new autotools broke the
following packages:

16:17:23: #1: Failed coreutils
16:18:14: #4: Failed file
16:18:35: #4: Failed flex
16:19:43: #1: Failed gdbm
16:21:13: #3: Failed gmp
16:28:33: #2: Failed libnet
16:35:34: #1: Failed mutt
16:42:28: #3: Failed rpm
16:43:02: #2: Failed shadow-utils

So far I managed to update all of them except for coreutils, mutt,
and shadow-utils (of course preserving all our patches if applicable).
The list of updated packages grew significantly during the last week.

Most of the time I've spent on rpm 4.11.2.  It looks much better than
our 4.2 (has a testsuite to verify its build results). There are still
some items I need to address before I consider the rpm package to be
ready to be committed, but so far it's looking good (was able to replace
4.2 with 4.11.2 on my system, now working on the upgrade procedure).

The reason why I'm going for package updates instead of patching and
fixing the existing packages is that I noticed that for most of our
userland we didn't have updates for many years.  Previously we were
struggling with packages switching to newer autotools.  Now the
autotools are quite stable (and most of the current releases use new
autotools), so the next challenge we currently facing is that more and
more packages rely on newer glibc's functions.

I'd love to join forces with somebody to push forward and do the glibc
update for Owl.  I need at least somebody to do x86_64 builds since
I simply don't have enough power to run x86/x_86_64 builds in parallel.

Right now I'm working on shadow-utils 4.2.1.  It's a major leap from
our (10 years old!) 4.0.4.1, so I expect that it would take me
considerable time to review the sources and re-apply our patches there.

I started to work on coreutils 8.22 also.  That work is almost completed
with the final Owl touches left to be applied.

My estimate is that I'll start unloading the results of my work on the
weekend.  By the amount of changes I've done so far, I'm wondering
whether we should look toward Owl 3.1 release? :)  If you agree, maybe
it would be safer to create the 3.1 branch (privately at this stage), so
all my work will go there.  Once it's ready we may announce it :).

-- 
(GM)

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ