1 |
On Sun, 9 Dec 2018 at 16:46, Philip Webb <purslow@××××××××.net> wrote: |
2 |
> |
3 |
> 181209 Marc Joliet wrote: |
4 |
> > Am Sonntag, 9. Dezember 2018, 11:35:16 CET schrieb Philip Webb: |
5 |
> >> What exactly are the "security reasons" ? |
6 |
> >> Do they apply to a single-user system ? -- if not, |
7 |
> >> why is the restrictive version of the policy file installed by default |
8 |
> >> rather than a warning at the end of the emerge output ? |
9 |
> > Good question. Checking the git log, the change was mode over two commits: |
10 |
> > https://gitweb.gentoo.org/repo/gentoo.git/commit/? |
11 |
> > id=02765dfc333e578af9e3fd525fc0067dc47d6528 |
12 |
> > https://gitweb.gentoo.org/repo/gentoo.git/commit/? |
13 |
> > id=df7afbda6b12a68578833225e694cee011b20342 |
14 |
> > The commit messages point to https://www.kb.cert.org/vuls/id/332928/ |
15 |
> > and https://bugs.gentoo.org/664236, |
16 |
> > which basically explain in more detail what Mick summarized yesterday. |
17 |
> |
18 |
> It looks to me like an over-reaction to a fairly unlikely exploit. |
19 |
> You are protected if you don't download images from untrusted sites |
20 |
> or if you don't run Ghostscript as root (who would ? ). |
21 |
> |
22 |
> It's true that you can use 'img2pdf' instead, which is perhaps the solution. |
23 |
|
24 |
More important than that, it seems the vulnerability is in |
25 |
ghostscript, and the vulnerable versions are not any longer even in |
26 |
portage, so shouldn't the change have been reverted by now? |
27 |
|
28 |
Arve |