Gentoo Archives: gentoo-portage-dev

From: Marius Mauch <genone@g.o>
To: gentoo-dev@g.o
Cc: drobbins@g.o, cycloon@×××××××.org, gentoo-portage-dev@g.o
Subject: [gentoo-portage-dev] GLEP 14 progress report
Date: Tue, 21 Oct 2003 19:59:39
Message-Id: 20031021210214.468920a0.genone@gentoo.org
1 Hi all,
2
3 this is a little update about the current progress of GLEP 14 (the
4 portage GLSA integration).
5 A few URLs first (most of the stuff is also somewhere in gentoo cvs):
6
7 The GLEP itself:
8 http://www.gentoo.org/proj/en/glep/glep-0014.html
9
10 The proposed new release plan for GLSAs:
11 http://gentoo.devel-net.org/glsa/release-plan
12
13 The DTD for future GLSAs (once the system is in place):
14 http://gentoo.devel-net.org/glsa/glsa.dtd
15
16 The userside code (I'm currently updating it and will commit it to cvs
17 this week, still have to figure out a few things):
18 http://gentoo.devel-net.org/glsa/glsa.py and glsa-check.py
19
20 Also plasmaroo has written a QT based editor which resides in the
21 gentoo-projects repository, and I've written the beginning of the GLSA
22 posting script.
23
24
25 If all goes well I think we can enter a first testing stage in 2 or 3
26 weeks, depend how much time I can donate to this project. If anyone has
27 a problem with the current plans please tell me ASAP so we can try to
28 fix it before the system is going live (so far the feedback has been
29 very positive).
30 Ok, so what still needs to be done before it can go live?
31
32 - the code has to be finished: mostly minor things as proper error
33 checking, output formatting and stuff like that. I don't expect big
34 problems here.
35
36 - infrastructure setup: as outlined in the release plan there is not
37 much to do from the infrastructure side, mostly setting up the right
38 directories and permissions.
39
40 - documentation: this is currently lacking, mostly because the code is
41 still work in progress. While docs are important I think it has to be
42 postponed until testing stage.
43
44 - portage integration: I think people have very different opinions on
45 this subject, so I'll just present my "vision":
46 I suggest we do a two-phase rollout. In the first phase the complete
47 code will be separate from the core portage code, so no direct
48 integration in emerge. People can use the seperate glsa-check script to
49 use the new system.
50 In the second phase we will integrate the basic functionalities of
51 glsa-check into emerge. Once that step is complete we will have a new
52 "security" package class beside system and world. There are other
53 features we could add like a security upgrade indicator or a security
54 update notification on emerge sync.
55
56 There were some concerns that the xml base for this project would
57 introduce xml code in portage. That's true to some degree: glsa.py which
58 contains all the backend code for handling the new GLSA format uses the
59 xml.dom.minidom python module for the parsing. So there is an indirect
60 import once we add GLSA support to emerge, but the portage code itself
61 doesn't need any xml code.
62 Also that module is part of the python package, so no additional
63 dependencies are added. This leads to another point. To make the system
64 secure we need gpg for the signature checking. So either we have to add
65 gpg as a (optional) dependency to portage or do some runtime checks, any
66 opinions on that issue?
67
68 If you have further questions/comments not adressed in this mail please
69 join #gentoo-security on irc.freenode.net and lets discuss it there or
70 reply to this mail on the gentoo-dev list.
71
72 Marius
73
74 --
75 Public Key at http://www.genone.de/info/gpg-key.pub
76
77 In the beginning, there was nothing. And God said, 'Let there be
78 Light.' And there was still nothing, but you could see a bit better.