Gentoo Archives: gentoo-dev

From: Daniel Campbell <contact@××××××××.us>
To: gentoo-dev@l.g.o
Subject: Re: [gentoo-dev] Policies for games dirs, new group "gamestat" for sgid binaries
Date: Sun, 22 Feb 2015 02:35:41
Message-Id: 54E94073.40609@sporkbox.us
In Reply to: Re: [gentoo-dev] Policies for games dirs, new group "gamestat" for sgid binaries by Ulrich Mueller
1 -----BEGIN PGP SIGNED MESSAGE-----
2 Hash: SHA256
3
4 On 02/21/2015 01:35 AM, Ulrich Mueller wrote:
5 >>>>>> On Fri, 20 Feb 2015, Daniel Campbell wrote:
6 >
7 >> When this becomes more widespread, what action are users urged
8 >> to take in order to "migrate" to the new system? Should our
9 >> everyday user account be removed from the `games` group, and the
10 >> group should be removed altogether?
11 >
12 > Currently, users need not take any action.
13 >
14 > In the hypothetical case that games.eclass would be abandoned, the
15 > "games" group would likely go away and should then be removed from
16 > users' systems. However, with about 1000 ebuilds currently
17 > inheriting games.eclass, I don't see that happening any time soon.
18 > There's a long discussion on this topic in the nethack bug [1].
19
20 Okay, I'll check that out. It seems that if the group is ever deemed
21 obsolete, it would gain a news item or something similar.
22
23 >
24 > Personally, I think that controlling who is allowed to run certain
25 > types of applications via group membership is a great idea. We
26 > should introduce that approach for other applications too. How
27 > about an "editors" group? Text editors are potentially dangerous
28 > because they allow users to modify files. Therefore, the system
29 > administrator should add only trusted users to the "editors" group
30 > so they can run programs like emacs, nano, or vim from the
31 > app-editors category.
32 >
33 > Ulrich
34 >
35 > [1] https://bugs.gentoo.org/125902
36 >
37 I hadn't thought of that! Would testing that idea require much beyond
38 creating the group, adding users, and chmodding the binaries? It seems
39 like it'd make a good USE option for those running servers with strict
40 permission needs. Then again, isn't that what LDAP or ACL are designed
41 to handle?
42 -----BEGIN PGP SIGNATURE-----
43 Version: GnuPG v2
44
45 iQEcBAEBCAAGBQJU6UByAAoJEJUrb08JgYgH80wH/RSeCKr8xF4N0WsY4mgfjtRX
46 TOWSJI/SF+Cb38IHUupKLrP0Hyp1VQud9lI2KZTS6UVj/qUhJBdOHMT9TGeAzZvI
47 rVss1RQMcW/ZPNhBa0M5i/mgopKqWaFyoLY9BWEqBa1cg26Sh43PbeFpdpI7wChR
48 4ya3NO9Hzc2zLLlpMjMGCVsJsuM7E24YonD9poGYP3LmBw4ZGnNN00YWnLofFMlj
49 8M3tr0FZ7ALlAsB2sb7vMTkUSX4s3t442Li+D1ihocZMGMSQ+NeJOwhzQAMAU7nx
50 3kzwGg+IDaaYrDOL4FnuJ4tOtc8DGKlzLj71VvJnhsvkxh2Hn3ljUZYeaJqoY1A=
51 =7xED
52 -----END PGP SIGNATURE-----

Replies