1 |
> |
2 |
> Wouldn't it be easier to do with a frozen -release tree? (In case you're |
3 |
> not getting it, I want a frozen tree... *grin*) |
4 |
|
5 |
----- |
6 |
|
7 |
I totally agree with Chris. It's difficult for software providers to support |
8 |
Gentoo because it's a moving target. |
9 |
|
10 |
I have seen two statements about a frozen tree; |
11 |
1 - administrators can use there own QA approved internal tree; |
12 |
this is true, however this takes a lot of effort to maintain those trees. They |
13 |
need to keep up with GLSA's and backport necessary ebuilds, digest, patches, |
14 |
etc. to their tree. But most importantly software providers don't know about |
15 |
these trees and are not able to QA their software on that tree. |
16 |
|
17 |
2 - it takes a lot of resources to maintain those trees; this is true if your |
18 |
goal is to provide ultimate stable and secure trees. Then it takes indeed a |
19 |
lot of effort because you need to test changes against every tree. There are |
20 |
other less ambitious scenarios possible; |
21 |
- just branch in cvs (2004.x) and do nothing with this branch. You have at |
22 |
least accomplished one big thing and that's predictability. It takes no extra |
23 |
resources at all to maintain these branches. Software providers can test |
24 |
their products against a tree and say their products works fine on Gentoo |
25 |
<release>. |
26 |
- create a branch in cvs and develop a tool that automatically backport GLSA's |
27 |
in the trees. This may not give us ultimate stable trees but at least we have |
28 |
secure and predictable trees. |
29 |
|
30 |
I agree that Gentoo Enterprise needs to form in small steps. I believe |
31 |
creating frozen trees takes very small amount of effort but gives us huge |
32 |
gains we can build on in the future. |
33 |
|
34 |
|
35 |
I suggest to take the following steps; |
36 |
- create frozen trees (gain predictability) |
37 |
- make Gentoo (portage) tree aware |
38 |
- extend GLSA tool to automatically backport GLSA's and their fixes to the |
39 |
trees |
40 |
- if financial and/or developer resources become available in the future |
41 |
backport stability fixes in the trees. |
42 |
|
43 |
-- |
44 |
gentoo-dev@g.o mailing list |