1 |
On 19.06.2019 18:45, William Hubbs wrote: |
2 |
> On Wed, Jun 19, 2019 at 09:24:10AM +0300, Mikle Kolyada wrote: |
3 |
>> We must keep gentoo as stable as possible taking care of our production |
4 |
>> users, I can speak of myself here, I am currently maintaining |
5 |
>> computation clusters with > 1k units (all are under gentoo), and I would |
6 |
>> not happy to see my cluster somehow affected by unstable features in the |
7 |
>> mainline (neither would others I believe). |
8 |
>> |
9 |
>> Still gentoo is the only meta distribution with freedom of choice, that |
10 |
>> says we can really implement bleeding edge solutions with a separate |
11 |
>> profile and something like "I_KNOW_WHAT_I_AM_DOING=1" variable set in |
12 |
>> the make.conf, the only question of implementation. |
13 |
> As I said in my last message on this thread, the stable tree is supposed |
14 |
> to be for that, so if your cluster is using fully stable packages I |
15 |
> would be concerned about this. |
16 |
The question implied nothing about stability of the existing tree, so I |
17 |
decided to emphasize (just in case) that stable state should be kept |
18 |
whatever we do. Some people just get paid for gentoo at work, that is |
19 |
where I was concerned. |
20 |
|
21 |
|
22 |
> |
23 |
> Is that what you are doing, or are you using ~ keywords for any of the |
24 |
> packages? |
25 |
|
26 |
Some gentoo packages I use simply do not have stable keywords, that is |
27 |
where I am using testing ones sometimes. |
28 |
|
29 |
Normally, as a member of (almost) all arch teams I push testing packages |
30 |
ahead and mark them stable once I am sure there is no bugs. |
31 |
|
32 |
|
33 |
> |
34 |
> Willia |