1 |
On Tue, 2006-11-28 at 14:56 -0600, James Potts wrote: |
2 |
> This looks good on the surface, Chris, but what happens in the case |
3 |
> where somebody wants to use the Release tree, but also wants (or |
4 |
> needs) one or more packages from the Live tree, and doesn't want to |
5 |
> switch completely over to the live tree? If I understand what you |
6 |
> want to do correctly, the Release tree would include only stable |
7 |
> packages. Other packages wouldn't just be masked, they would be |
8 |
> completely unavailable to anybody using that tree. |
9 |
|
10 |
Correct. The whole concept of the release tree is it is a complete |
11 |
package. All of the parts are supposed to work together. There's |
12 |
*nothing* that is not considered stable. If you need other packages, |
13 |
you add them to an overlay and support them yourself. |
14 |
|
15 |
> I like the idea of having a stable p.mask much better, which says |
16 |
> "profile" to me. Any thoughts on a special profile just for releases? |
17 |
|
18 |
It is completely unworkable with the 11,000+ packages we currently have |
19 |
in the tree, whereas an automated script that parses the current tree is |
20 |
much simpler. Quite simply, when GLEP19 was being worked on, it became |
21 |
painfully obvious almost immediately that even with the 10+ people who |
22 |
volunteered to work on it, that it would be impossible to maintain a |
23 |
profile-based system. We estimated that it would take us more than |
24 |
twice as many developers working on *only* the stable tree than we had |
25 |
working on the "live" tree to even keep up. |
26 |
|
27 |
To put it simply, I have *zero* interest in any profile/mask-based |
28 |
concepts for providing "stable" as they don't work. All they do is |
29 |
create an enormous bottleneck and a massive amount of workload for every |
30 |
single developer. With the release trees, essentially only those |
31 |
interested in supporting the tree are required to work on it. The tree |
32 |
is created entirely by scripts and is tested *before* it's released on |
33 |
the public. |
34 |
|
35 |
-- |
36 |
Chris Gianelloni |
37 |
Release Engineering Strategic Lead |
38 |
Alpha/AMD64/x86 Architecture Teams |
39 |
Games Developer/Council Member/Foundation Trustee |
40 |
Gentoo Foundation |