1 |
On 8/20/13 11:19 AM, William Hubbs wrote: |
2 |
> During the last release of OpenRC, I learned that people *do* run |
3 |
> production servers on ~arch. I asked about it and was told that the |
4 |
> reason for this is bitrot in the stable tree. |
5 |
|
6 |
People frequently point to lack of manpower as reason for this, but I |
7 |
don't think it's true. |
8 |
|
9 |
First, the rate arch teams (at least mainstream, amd64/x86) deal with |
10 |
stabilization queues has improved. I'm not aware of any severe backlogs |
11 |
in terms of stabilization bugs filed but not handled (on amd64/x86). |
12 |
|
13 |
Second: people maintain list of packages to unmask, or maintain ~arch |
14 |
servers... I hope they have some kind of testing/staging environment |
15 |
before unleashing ~arch on production servers. Well, this is exactly |
16 |
what stable can be - this effort could easily be shared with the rest of |
17 |
the community by these people stabilizing things. I'm in favor of policy |
18 |
changes if needed to make this happen. |
19 |
|
20 |
Furthermore, it'd be great to have people who actually run servers help |
21 |
with stabilizations. Arch teams could be doing their best, but they |
22 |
probably don't run postgres/apache/ruby/whatever in production, and so |
23 |
it's pretty much if it compiles and reverse dependencies don't explode |
24 |
it's perfect. People running servers in production can actually test |
25 |
these pretty thoroughly in staging environment and either vet them with |
26 |
higher confidence or file good blocking bugs. |
27 |
|
28 |
Paweł |