1 |
On Tuesday 03 February 2004 13:19, Kevyn Shortell wrote: |
2 |
> A lot of ebuilds have been 'accidently' deleted by over-zealous space |
3 |
> reclaimation workers not knowing which builds work on all supported |
4 |
> arches. |
5 |
> |
6 |
> Yes a stable branch is a lot of work. But that's the whole point. It is |
7 |
> a 'stable' branch thereby implying a lot of work has gone into it to |
8 |
> make it so. Until we can go 6 months without someone deleting a critical |
9 |
> ebuild that's required by an alt-arch just because they their particular |
10 |
> arch has a newer version, we need a seperate branch. |
11 |
|
12 |
If that's the case, we won't be helping fix the problem, we'll just have a |
13 |
hopefully never-broken stable tree and a sometimes-broken main tree. |
14 |
|
15 |
I don't know the details of the accidental deletions. In fact not having |
16 |
committed to Gentoo cvs in about 7 months (though hoping to do so again |
17 |
sometime soon) I don't know how much e.g. repoman can help in such cases. Is |
18 |
it a really insoluble problem with the keywords system and current tools, or |
19 |
just developers who aren't careful/thorough enough? |
20 |
|
21 |
In either case, stable-marked ebuilds shouldn't only ever be removed at their |
22 |
set 'timeout' dates every quarter, at which point the entire tree is |
23 |
rechecked for deps etc. anyway. As for other modifications, if an ebuild is |
24 |
marked stable for one arch but not for another, any change to it that's not |
25 |
meant to affect the stable tree needs a new revision. Is it too much to ask |
26 |
that developers adhere to rules like these? |
27 |
|
28 |
I can understand a separate stable tree if we expect that stable and |
29 |
non-stable keywords will rarely be given to the same ebuild revision, thus |
30 |
eliminating the single-tree benefit of code sharing/reuse. That would however |
31 |
imply that many ebuilds will be revision-bumped (not version-bumped) even |
32 |
though they've existed for a long time (to enter the stable tree) and a newer |
33 |
version is usually available. Does someone have statistics or predictions for |
34 |
this? |
35 |
|
36 |
> |
37 |
> > About keyword naming, I agree with Stuart's note elsewhere in the thread |
38 |
> > that 'stable' is misleading. I also want to ask how the transition of |
39 |
> > arch-->~stable:arch-->stable:arch is different from the existing |
40 |
> > transition of ~arch-->arch. |
41 |
> |
42 |
> From an enterprise perspective having something tagged stable after 2 |
43 |
> weeks, does not qualify as stable, that's enough to move it to the alpha |
44 |
> stage. Having something beat on for a month with no critical bugs might |
45 |
> qualify for the beta stage, have it go for 2 months without any |
46 |
> significant bugs might qualify for the stable monkier. Just because it |
47 |
> doesn't have any obvious bugs doesn't mean there isn't any. Memory leaks |
48 |
> come to mind. Those take days, weeks sometimes to be flushed out. |
49 |
> |
50 |
> > If it isn't different and it's just a matter of the package being more |
51 |
> > and more tested and used and proven without known (unfixed) |
52 |
> > bugs/vulnerabilites, I don't think it's appropriate to create keywords by |
53 |
> > adding several modifiers to an arch's name (~ and stable). We're not |
54 |
> > really combining the properties of ~ and 'stable', and might as well |
55 |
> > assign stability levels with keywords like 0:x86 for ~x86, ..., 3:x86 for |
56 |
> > stable:x86. |
57 |
> |
58 |
> I think this is where having a seperate branch would actually help. If |
59 |
> it's in the stable branch, there's no keyword required to tag it as |
60 |
> such. |
61 |
|
62 |
I wasn't complaining about introducing extra keywords, just that they weren't |
63 |
consistent with existing ones. As I understand you, you agree that the new |
64 |
keywords simply represent a gradual increase in known/perceived package |
65 |
stability and testedness, a difference in quality and not in quantity. Is |
66 |
this correct? |
67 |
|
68 |
> Just goes to show that something in heavy use, with no bug reports can't |
69 |
> really always be considered as stable, just because you haven't seen any |
70 |
> issues. If we're going to have a real 'stable' profile/branch/whatever |
71 |
> it needs to have serious testing and accountability before we slap the |
72 |
> stable monkier on it. Our reputation depends on it. |
73 |
|
74 |
Yes. However, the GLEP and what's been said in the thread so far hasn't |
75 |
addressed any (formal?) extra testing etc. to be performed by us, only |
76 |
judging that a package is stable. Is this still going to be left at the |
77 |
maintainer's discretion? |
78 |
|
79 |
-- |
80 |
Dan Armak |
81 |
Gentoo Linux developer (KDE) |
82 |
Matan, Israel |
83 |
Public GPG key: http://dev.gentoo.org/~danarmak/danarmak-gpg-public.key |
84 |
Fingerprint: DD70 DBF9 E3D4 6CB9 2FDD 0069 508D 9143 8D5F 8951 |