1 |
On Sun, 08 Feb 2009 14:43:01 -0800 |
2 |
Zac Medico <zmedico@g.o> wrote: |
3 |
> Well, if you want to use timestamps, the alternative is to |
4 |
> distributors to use a protocol which preserves timestamps. This |
5 |
> creates an unnecessary burden. Allowing distribution of metadata |
6 |
> cache via version control systems is more flexible. |
7 |
|
8 |
Ok, if we're going to encourage this, let's do it properly: |
9 |
|
10 |
* Have a branch called 'master'. Commit to it. Don't stick any metadata |
11 |
in it. |
12 |
|
13 |
* Have a branch called 'master-with-metadata'. Don't commit to it |
14 |
manually. |
15 |
|
16 |
* Have a script that merges master to master-with-metadata, and as part |
17 |
of the merge commit, generates all necessary metadata for the range |
18 |
it's merging. |
19 |
|
20 |
* Store either the partial hash or the owning repository and timestamp |
21 |
of each eclass used by an ebuild in its metadata. |
22 |
|
23 |
-- |
24 |
Ciaran McCreesh |