1 |
On Fri, Dec 03, 2004 at 08:31:39AM -0500 or thereabouts, Chris Gianelloni wrote: |
2 |
> The only problem that I see with this is that he has come up with no |
3 |
> real viable solution other than offering a stage3 tarball. That is fine |
4 |
> and dandy, but Release Engineering is working on the future removal of |
5 |
> all tarballs *except* for stage1. The other stageballs would be created |
6 |
> on-the-fly. Now, this won't be the case for 2005.0 or even 2005.1, but |
7 |
> I suspect by 2006.0 we will not be offering any stages other thana |
8 |
> stage1. |
9 |
|
10 |
As Jason said in a previous email, portage should be fixed to handle |
11 |
circular dependency stuff around 2005.1. I would think this alleviates a |
12 |
lot of the concerns that rac has. (though I don't pretend to understand |
13 |
the issues to the same level that either Jason or Rob do) |
14 |
|
15 |
> What this means is if we have something wrong with a stage1 tarball |
16 |
> that: |
17 |
> |
18 |
> #1. rac should talk to releng rather than posting contrary information |
19 |
|
20 |
To be fair, when you and I talked in #gentoo-installer a few weeks ago |
21 |
about dumping stage tarballs, I told you that rac had some concerns about |
22 |
this and that you should talk to him about it. |
23 |
|
24 |
> #2. we should work to resolve any problems with the stage1 tarball |
25 |
|
26 |
I don't think anyone disagrees with this. I think rac's solution is more |
27 |
of a band-aid than anything. I think he along with everyone else would |
28 |
like to see a better, more robust solution put in place. In the mean time, |
29 |
however, I don't think it hurts to notify our users of the problems with |
30 |
stage 1 tarballs, *in their current form*, and suggest a work-around while |
31 |
we work to fix them for a future version. |
32 |
|
33 |
--kurt |