Gentoo Archives: gentoo-dev

From: Marius Mauch <genone@g.o>
To: gentoo-dev@g.o
Subject: Re: [gentoo-dev] GLEP 14 follow-up / security project
Date: Wed, 24 Sep 2003 04:01:17
Message-Id: 20030924060104.4fa60c27.genone@gentoo.org
In Reply to: Re: [gentoo-dev] GLEP 14 follow-up / security project by Grant Goodyear
1 On 09/23/03 Grant Goodyear wrote:
2
3 > I'm glad to see this project underway. I do have a couple of
4 > questions, though.
5 >
6 > > Security bugs should be kept in bugzilla,
7 >
8 > What's the rationale for keeping GLSA's in bugzilla, since bugzilla is
9 > going to have to be hacked to make it work?
10
11 As other people are also against this we should drop it. I just picked
12 that up from the previous discussion in #gentoo-security. So we make a
13 separate tracking system, not sure if it will be easier or harder to do,
14 but definitely a cleaner approach.
15
16 > > The actual filing and editing of these bugs should be done with a
17 > > new interface that is specially designed for security bugs and GLSA
18 > > information. Once a security bug is marked as fixed a GLSA
19 > > generation script is run that generates the GLSA, GPG-signs it
20 > > (depending on policy) and distributes it on mailing lists, http- and
21 > > rsync-servers.
22 >
23 > Where is this script run? If it's on a gentoo server, then I don't
24 > really like the idea of the script signing the GLSA. Perhaps I'm just
25 > being paranoid, but I would really prefer the signing to be performed
26 > by the user issuing the GLSA. If, on the other hand, there is a GLSA
27 > tool that devs can run on their own machines that assists the dev in
28 > creating the GLSA, then signs the GLSA and uploads it to the
29 > appropriate location, that would be just fine with me.
30
31 This should be decided by the security project, there are other people
32 who can comment/decide that better than me.
33
34 > > The update script then can take the GLSAs from /usr/portage/glsa or
35 > > the http repository (to avoid unneeded syncs just to get the GLSA).
36 >
37 > Drobbins has said that he would prefer the update script to be
38 > incorporated into emerge as soon as possible. I get the impression
39 > from Carpaski that we can, indeed, do that.
40
41 I'm not against portage integration, but I'd rather want to finish and
42 test the script and the GLSA framework before we put it into emerge. We
43 wouldn't do anyone a favor if we put unfinished stuff in portage. The
44 integration isn't that difficult, just adding some files to the portage
45 tarball and changing a few calls to use portage functions instead of
46 emerge / portageq.
47
48 > Nice job!
49 > -g2boojum-
50
51 Thanks
52
53 Marius
54
55 --
56 Public Key at http://www.genone.de/info/gpg-key.pub
57
58 In the beginning, there was nothing. And God said, 'Let there be
59 Light.' And there was still nothing, but you could see a bit better.