1 |
On 02/04/14 21:22, Mike Gilbert wrote: |
2 |
> On Wed, Apr 2, 2014 at 12:52 PM, Samuli Suominen <ssuominen@g.o> wrote: |
3 |
>> The "30 days maintainer time out" stabilization policy isn't working |
4 |
>> when package has multiple SLOTs, because |
5 |
>> the bugs are filed for only latest SLOT, where as some packages require |
6 |
>> stabilization in sync at both SLOTs |
7 |
>> |
8 |
>> Option 1: |
9 |
>> |
10 |
>> Either revert the whole policy, and never CC arches on unanswered bugs |
11 |
>> when the package has a maintainer, |
12 |
>> and let him do it when he finds the time himself, and if that doesn't |
13 |
>> happen, wait until it's dropped to maintainer-needed@ |
14 |
>> |
15 |
>> Option 2: |
16 |
>> |
17 |
>> Or, the person who is CCing the arches in 30 days timeout, needs to make |
18 |
>> sure the bug covers all SLOT at the same time |
19 |
>> |
20 |
>> |
21 |
>> The status quo no longer allows me to maintain stable version of |
22 |
>> dev-libs/girara, app-text/zathura*, and the issue needs |
23 |
>> to be addressed, see http://bugs.gentoo.org/502714 for what inspired |
24 |
>> this mail |
25 |
>> |
26 |
>> - Samuli |
27 |
>> |
28 |
> If you want to prevent packages from being "timed out", just leave a |
29 |
|
30 |
That would require me to take action on the STABLEREQ, which I don't |
31 |
want to do |
32 |
if I see someone requesting it stable only because it's newer version, |
33 |
see below: |
34 |
|
35 |
> comment on the bug saying so. If you don't even have time to do that |
36 |
> within a 30 day window, why are you the maintainer? |
37 |
|
38 |
It's like we handle stabilization for Xfce, we collect enough xfce-base/ |
39 |
and xfce-extra/ |
40 |
packages together to form a longer list, to avoid bothering arch teams |
41 |
unnecessarily |
42 |
constantly |
43 |
This time it's app-text/zathura-meta and it's dependencies, multiple |
44 |
packages, we'd |
45 |
like to wait there are enough plugins, and form a list |
46 |
|
47 |
And this time the package even had 3 maintainers, 2 gentoo developers |
48 |
and one proxy, |
49 |
in metadata.xml, yet none of us managed to intervene in time to prevent |
50 |
the SLOT breakage[1] |
51 |
done by arches without our consent |
52 |
|
53 |
[1] Stabilizing headers, but not the library, causing the .h not to |
54 |
match the .so, causing every |
55 |
reverse dependency of the library to break |
56 |
|
57 |
> |
58 |
> Another option would be to add some kind of notation to metadata.xml. |
59 |
> |
60 |
|
61 |
That could work |