[OmniOS-discuss] Bug ?

[Aurélien Larcher]

On 03/07/2017 16:41, Jim Klimov wrote:
> On July 1, 2017 1:02:30 AM GMT+02:00, "Aurélien Larcher" <aurelien.larcher at gmail.com> wrote:
>> I guess one should remind the little elves of the magical forest that
>> they need to take over the maintenance, they were quite busy with
>> Midsummer apparently.
>>
>> Otherwise, basically Guenther's advice is the way to go:
>> - for illumos specific issues: https://www.illumos.org/issues,
>> - for software package bugs we could look at it in OpenIndiana and
>> someone could backport the fix to OmniOS afterwards (when
>> infrastructure is fixed).
>>
>> I hoped that we could have a virtuous relationship with OpenIndiana
>> Hipster acting as a "bloody" and OmniOS continuing as a stable subset,
>> similarly to the current scheme (also benefiting from the current
>> Hipster CI setup to avoid losing momentum).
>> There is apparently no interest following the discussion which was
>> initiated right after the announcement, this is a bit disappointing.
>>
>> I hope that people will join forces to consolidate the landscape of
>> community-supported illumos distros.
>> At some point one needs to stop talking theory and tragedy, there are
>> pragmatic solutions to implement: there are infrastructures in place
>> in both OI and OmniOS-land, there are overall directions and goals in
>> both (rolling dev vs stable), there is a solid release engineering
>> practice in OmniOS, it does not seem like we swim in complete
>> abstraction.
>> Contributing software packages can be easy, the difficulty seems to
>> get people to figure out that maintaining even one package is a
>> meaningful contribution.
>>
>> (I maintain a bunch of packages for OpenIndiana and I do not work in
>> IT)
>>
>> Kind regards,
>>
>> À Jeudi 29 juin 2017, Davide Poletto a écrit :
>>> The more I think *how fast* that's (supposedly) happening the more I
>> feel
>>> shocked.
>>>
>>> On Jun 29, 2017 10:33 PM, "Floris van Essen ..:: House of Ancients
>> Amstafs
>>> ::.." <info at houseofancients.nl> wrote:
>>>
>>> Hi All,
>>>
>>> With pain in my heart, i fear OmniOs has started it's slow descent to
>> death
>>> :-(
>>>
>>>
>>>
>>> -----Oorspronkelijk bericht-----
>>> Van: OmniOS-discuss [mailto:omnios-discuss-bounces at lists.omniti.com]
>> Namens
>>> Paul B. Henson
>>> Verzonden: donderdag 29 juni 2017 22:13
>>> Aan: 'Guenther Alka' <alka at hfg-gmuend.de>;
>> omnios-discuss at lists.omniti.com
>>> Onderwerp: Re: [OmniOS-discuss] Bug ?
>>>
>>>> From: Guenther Alka
>>>> Sent: Thursday, June 29, 2017 12:33 AM
>>>>
>>>> There are no signs from OmniTi to add any fixes.
>>> An OmniTI person posted an intention to back port security fixes to
>> r151022
>>> for one year, although I'm not sure whether that was under the
>> auspices of
>>> OmniTI employment or just on his personal time. So far though, there
>> have
>>> been no new commits in the r151022 repo since Dan's last one on
>> 5/10..
>>>
>>> _______________________________________________
>>> OmniOS-discuss mailing list
>>> OmniOS-discuss at lists.omniti.com
>>> http://lists.omniti.com/mailman/listinfo/omnios-discuss
>>> ...:: House of Ancients ::...
>>> American Staffordshire Terriers
>>>
>>> +31-628-161-350
>>> +31-614-198-389
>>> Het Perk 48
>>> 4903 RB
>>> Oosterhout
>>> Netherlands
>>> www.houseofancients.nl
>>> _______________________________________________
>>> OmniOS-discuss mailing list
>>> OmniOS-discuss at lists.omniti.com
>>> http://lists.omniti.com/mailman/listinfo/omnios-discuss
>>>
>> -- 
>> Thanks for sailing Jolla :)
>> _______________________________________________
>> OmniOS-discuss mailing list
>> OmniOS-discuss at lists.omniti.com
>> http://lists.omniti.com/mailman/listinfo/omnios-discuss
> Among technological differences between OI and OO, a notable benefit is LX zones. It was often stated that OI community has not enough resources to deviate from upstream illumos-gate, since maintaining a fork beyond a branding patch or so is indeed complicated.
I guess it is more a matter of priority: we had the userland migration
ongoing and updating/integrating quite many components in the process
was very time consuming.
Last time we discussed LX the conclusion was that we need one person
dedicated to the task, and otherwise it is pointless to push it out if
we cannot guarantee that version will be updated following a regular
schedule.
> I got wondering if it is possible to take omnios-gate and build just the LX-related bits for the sake of adding the LX brand support as a few packages installable to "vanilla" illumos-gate os/net? In the binary end, it does add just some modules, tools and scripts, right? At least, if this works, it might help until LX is upstreamed...
Last year I rebased illumos-omnios, added some stuff (iwn + loader) and
built it on Hipster just to see how it is done , but I have not had
enough interest otherwise.
I am more interested in compiler migration and parallel (scientific)
computing tools as you know.

I wonder what is the most time consuming?
1) sharing one userland and getting illumos-omnios to build on both
OmniOS and OI,
2) keeping illumos-omnios as it is and duplicate userland.

Beyond the maintenance the question is: how do you provide a stable
system without a testing branch?
If everybody is interested in using a stable OmniOS, who is developing
and testing "Bloody"?

For instance:
- who will take care of the migration to recent GCC?
- will this task be a duplicate of what is done in oi-userland or will
there be some level of collaboration?
- if the tools are different, how should this be implemented?

These questions should be addressed as soon as possible to facilitate
any transition that may be required.
I understand that it is tempting to keep things as they are, but one
should consider if the direction is viable in the long term.
Maintaining the current r22 is one thing, preparing the next cycle is
another problem: if goals and timeline are not set according to
available resources, one may end up with nothing.

Again, I am not directly involved and my point of view is very naive, I
am just interested in people thoughts on the topic.
Usually it *just* starts with somebody taking the initiative, sit and
deal with the code. ;)

>
> Jim
> --
> Typos courtesy of K-9 Mail on my Redmi Android