1 |
On 24-08-2011 09:20:00 +0000, Robin H. Johnson wrote: |
2 |
> On Wed, Aug 24, 2011 at 11:01:07AM +0200, Fabian Groffen wrote: |
3 |
> [snip] |
4 |
> > I would like to put an additional issue on the table which I encountered |
5 |
> > while implementing fully from CVS generated ChangeLogs for the Prefix |
6 |
> > rsync tree. |
7 |
> [snip] |
8 |
> And all of this is impacted by how we do thin manifests and commit |
9 |
> signing. |
10 |
|
11 |
Obviously. |
12 |
|
13 |
> Thin manifests are those that contain ONLY entries for files not covered |
14 |
> by another (direct or indirect) hash in the VCS. Git's use of SHA1 |
15 |
> allows a conversion from Git+thin Manifest to classical Manifest2. |
16 |
|
17 |
How does this work with generating the other hashes? |
18 |
|
19 |
> Commit signing has a few implications/side-effects: |
20 |
> - commits are signed so Manifests are NOT signed anymore. |
21 |
|
22 |
I assume this is ok, and has no effect in terms of guarantees one makes |
23 |
about the content. |
24 |
|
25 |
> - During the conversion to classical Manifest2, we need to create |
26 |
> automated signatures (see the tree-signing GLEPs for MetaManifest). |
27 |
> - As a side advantage of the automated Manifests/signatures, we can use |
28 |
> the Manifest2 changes proposed in the tree-signing GLEPs to cover ALL |
29 |
> of the profiles and eclasses. |
30 |
> |
31 |
> I'm going to be away Thursday till Monday, so the lack of any further |
32 |
> impact from me doesn't mean I don't have an opinion, rather just that |
33 |
> I'm away from the Internet. |
34 |
|
35 |
Your input is much appreciated (if not authoritive). I hope you'll |
36 |
catch up on this discussion after you return. |
37 |
|
38 |
|
39 |
-- |
40 |
Fabian Groffen |
41 |
Gentoo on a different level |