1 |
On Tuesday 03 February 2004 15:41, Kurt Lieber wrote: |
2 |
> On Tue, Feb 03, 2004 at 12:04:59PM +0200 or thereabouts, Dan Armak wrote: |
3 |
> > A real separate cvs branch seems like a lot of extra work; most updates |
4 |
> > going into the stable branch will probably also go into the main tree. |
5 |
> > What am I missing? |
6 |
> |
7 |
> A key part of the GLEP is ensuring that ebuilds stay in the tree for a |
8 |
> minimum of one year. As has been proven time and time again, we don't have |
9 |
> the necessary QA or control over our current tree to offer this feature, so |
10 |
> I felt it was betetr implemented by offering a separate tree. |
11 |
|
12 |
I've just heard more details on irc about the ebuild deletion problem. I have |
13 |
to say it gives me a very unpleasant feeling that we can't trust our devs not |
14 |
to delete necessary files (in this case, any files with stable keywords, very |
15 |
easy to check). And we're also accepting the fact that the main tree gets |
16 |
broken in this way once in a while. That's bad, regardless of the stable tree |
17 |
issue. |
18 |
|
19 |
> |
20 |
> Also, with one tree, we cannot offer atomic updates to our users. That is |
21 |
> to say, with one tree, there would have to be some period of time where the |
22 |
> stable tree was 'in flux' as devs went through and marked new ebuilds |
23 |
> stable. With a separate tree, devs have an entire quarter where they can |
24 |
> mark things stable at their leisure. Then, at a pre-defined date, we pull |
25 |
> those ebuilds and can offer one, consistent tree to those folks who want |
26 |
> the stable branch. |
27 |
|
28 |
We could use ~stable for the flux period, and mark a bunch of ~stable stuff as |
29 |
stable atomically. |
30 |
|
31 |
> As for your question on transition, for 95% of all stable ebuilds, the path |
32 |
> should be: |
33 |
> |
34 |
> ~arch --> arch --> stable:arch |
35 |
> |
36 |
> ~stable is there primarily for off-cycle updates. If we need to issue a |
37 |
> GLSA and updated ebuild with very little testing, it would be included in |
38 |
> the stable tree marked as ~stable:arch. Then, after 'adequate' testing, it |
39 |
> would be moved to stable:arch |
40 |
|
41 |
~stable as staging/temp area makes sense, more so than yet another level of |
42 |
stability. Sounds good. |
43 |
|
44 |
> In terms of what constitutes 'stable', that will largely be left up to the |
45 |
> herd and/or package maintainer. Remember, 'stable' refers more to the fact |
46 |
> that the tree itself does not change, rather than the ebuilds themselves |
47 |
> being more stable. (This is another example of something that may change |
48 |
> down the road once we improve our QA efforts) |
49 |
|
50 |
On reading the GLEP again I see that it says just this. Looking at other |
51 |
messages in this thread, I don't think I was the only one who read into it |
52 |
that we need much improved QA efforts before the first stable tree release. I |
53 |
agree that a gradual transition is better. |
54 |
|
55 |
-- |
56 |
Dan Armak |
57 |
Gentoo Linux developer (KDE) |
58 |
Matan, Israel |
59 |
Public GPG key: http://dev.gentoo.org/~danarmak/danarmak-gpg-public.key |
60 |
Fingerprint: DD70 DBF9 E3D4 6CB9 2FDD 0069 508D 9143 8D5F 8951 |