1 |
On 10/08/17 11:42, William L. Thomson Jr. wrote: |
2 |
> On Thu, 10 Aug 2017 10:50:45 +1000 |
3 |
> "Sam Jorna (wraeth)" <wraeth@g.o> wrote: |
4 |
> |
5 |
>> On 10/08/17 06:35, William L. Thomson Jr. wrote: |
6 |
>>> FYI binpkgs have no hash. If someone did something malicious within |
7 |
>>> the binhost to the binpkgs. You have no way of knowing. Yes the |
8 |
>>> same can happen with ebuilds and manifest. But easy to sync portage |
9 |
>>> and see if a manifest has changed. |
10 |
>> |
11 |
>> This isn't exactly true - see ${PKGDIR}/Packages on the binhost, which |
12 |
>> is a manifest of built packages and related metadata. Granted this is |
13 |
>> created by the binhost, it does exist and contains SHA1 and MD5 |
14 |
>> hashes, as well as package size. In that sense it's no different to |
15 |
>> how a package Manifest file works within a repository. |
16 |
> |
17 |
> You are correct. I meant to say no verifiable hash. You can hash |
18 |
> anything locally and claim it to be trustworthy. Thus mentioning |
19 |
> syncing portage to compare manifest of ebuild/SRC_URI<snip> |
20 |
> IMHO SRI_URI is more trustworthy than binhost, in the sense of |
21 |
> verification. If you have means to verify the binhost stuff it maybe |
22 |
> more trustworthy. That is left to the admin. |
23 |
|
24 |
This is no greater risk than syncing from a potentially compromised |
25 |
mirror. You would use a mirror you trust and, similarly (perhaps even |
26 |
more so) you would use a binhost you trust. |
27 |
|
28 |
It does raise the idea of some form of signing of the Packages file, |
29 |
similar to gpg-signed portage snapshots, but that's moving well beyond |
30 |
the scope of this thread. |
31 |
|
32 |
-- |
33 |
Sam Jorna (wraeth) |
34 |
GnuPG ID: D6180C26 |