1 |
On Sun, 18 Apr 2010 18:25:02 +0530 |
2 |
Nirbheek Chauhan <nirbheek.chauhan@×××××.com> wrote: |
3 |
|
4 |
> ====== |
5 |
> Coding: |
6 |
> ====== |
7 |
> * Portage support for thin manifests |
8 |
> - Most important; a release must be done atleast 1 month before the |
9 |
> migration |
10 |
> - Testing can be done with overlays |
11 |
> * Repoman support for thin manifests |
12 |
> - Testing can be done with overlays |
13 |
|
14 |
These two are directly connected together. |
15 |
|
16 |
I've already coded some basic support for Thin Manifests in my |
17 |
newly-created Portage branch. The branch itself can be looked up at: |
18 |
http://git.mgorny.alt.pl/portage/log/?h=git-repo |
19 |
|
20 |
And the particular commit I'm mentioning here: |
21 |
http://git.mgorny.alt.pl/portage/commit/?h=git-repo&id=2b77f313b89b84dcc04299b83ef9ecfd73318cd6 |
22 |
|
23 |
I've did as much as I could without getting more information, and |
24 |
turning everything around. What it can is: |
25 |
1) Detect '.git' and enable Thin Manifests (only dirty check), |
26 |
2) generate Thin Manifests with only DIST checksums, |
27 |
3) merge ebuilds with Thin Manifests. |
28 |
|
29 |
I have tested it against a random kde-sunset and x11 overlay ebuilds |
30 |
(the former having DIST checksums, the latter being empty). What we |
31 |
don't (I guess so, I haven't checked it) support now is non-existent |
32 |
Manifests. Are we supposed to support them, or should we always |
33 |
generate Manifests, even if they will be empty? |
34 |
|
35 |
-- |
36 |
Best regards, |
37 |
Michał Górny |
38 |
|
39 |
<http://mgorny.alt.pl> |
40 |
<xmpp:mgorny@××××××.ru> |