| 1 |
On Sat, Nov 21, 2015 at 1:51 PM, Andrew Savchenko <bircoph@g.o> wrote: |
| 2 |
> |
| 3 |
> On Wed, 18 Nov 2015 07:01:21 -0500 Rich Freeman wrote: |
| 4 |
>> On Wed, Nov 18, 2015 at 6:12 AM, Alexander Berntsen <bernalex@g.o> wrote: |
| 5 |
>> > When I do QA in projects I'm involved with (at least outside of |
| 6 |
>> > Gentoo), we don't do it live on end-user systems. I'll leave the |
| 7 |
>> > details as an exercise for the Gentoo developer. |
| 8 |
>> > |
| 9 |
>> |
| 10 |
>> People who run ~arch are not really end-users - they're contributors |
| 11 |
>> who have volunteered to test packages. |
| 12 |
> |
| 13 |
> I strongly disagree with you. We do not use stable even at |
| 14 |
> enterprise grade production systems and HPC setups. Stable is just |
| 15 |
> too freaking old in order to be usable for our purposes, not to |
| 16 |
> mention that it lacks many packages at all. We tried stable |
| 17 |
> several times, it just freaks out admins (including myself) too |
| 18 |
> badly or results in horrible mess of stable and unstable which is |
| 19 |
> less stable that unstable setups. I do not use stable at |
| 20 |
> workstations and personal setups as well. |
| 21 |
|
| 22 |
Interesting. I've had the opposite experience, and don't run ~arch |
| 23 |
except for testing purposes. I don't hesitate to keyword packages |
| 24 |
when necessary, and file bugs for their stabilization if appropriate. |
| 25 |
|
| 26 |
Also, if you're doing something like HPC then you're probably focused |
| 27 |
on a specific application, with your own QA system, so Gentoo's QA |
| 28 |
doesn't really impact you much anyway as your own regression test is |
| 29 |
going to catch issues. I'm not nearly that formal but I've |
| 30 |
containerized almost all my services because I don't like relying on |
| 31 |
Gentoo's QA. If I update my mariadb container I just make sure that |
| 32 |
mariadb is working, and revert it if not. If it happens to contain a |
| 33 |
broken ssh client it doesn't concern me at all, since I don't use that |
| 34 |
container for ssh. Of course, the downside of this is that I end up |
| 35 |
updating a lot of hosts, all for personal use. |
| 36 |
|
| 37 |
> Of course I understand that there are people |
| 38 |
> using it and I try to support stable packages as well, but these |
| 39 |
> versions are mostly a burden and I can't really understand stable |
| 40 |
> users. |
| 41 |
|
| 42 |
Well, to be fair it seems like most Gentoo developers consider half |
| 43 |
the tree a burden (that would be the "other" half). We all have our |
| 44 |
itches that we're trying to scratch. As long as everybody follows the |
| 45 |
policies the results end up working out reasonably well for everybody. |
| 46 |
Some of us barely test ~arch at all, and others barely test stable at |
| 47 |
all, and it seems that for the most part things work out. |
| 48 |
|
| 49 |
In any case, the purpose of ~arch is testing, and is not intended to |
| 50 |
be a stable experience, even if it often ends up being that way (which |
| 51 |
is certainly nothing to complain about). If we added another layer of |
| 52 |
testing above ~arch, all we'd see happen is that everybody who runs |
| 53 |
~arch today would just switch to that, since it would essentially be |
| 54 |
the same thing, and ~arch wouldn't really serve any purpose at all. |
| 55 |
If the purpose of ~arch isn't testing, then why have it at all? |
| 56 |
|
| 57 |
But, like I said, if somebody wants to volunteer to do a barrage of QA |
| 58 |
tests on portage, by all means do so. It will only make life better |
| 59 |
for everybody. I just don't see any reason to bar the portage authors |
| 60 |
from introducing a version if they consider it suitable for testing. |
| 61 |
|
| 62 |
-- |
| 63 |
Rich |
Archives