[OpenIndiana-discuss] OI roadmap (for production)

[Nikola M]

On 12/ 4/15 04:01 PM, Alexander Pyhalov wrote:
> On 12/04/2015 17:16, Dmitry Kozhinov wrote:
>> Alexander Pyhalov, Thu Dec 3 12:39:35 UTC 2015:
>>  > As for /dev, I think it's not a big secret that it's efficiently 
>> dead.
>>
>> Nikola M, Thu Dec 3 18:46:38 UTC 2015:
>>  > It is not big secret that hipster is never planned for any kind of
>> production
>>
>> What is the plan then?
>
> Hi.
> What do you want to know?
> Currently we try to ride the car and tune it simultaneously (like 
> illumos-gate does). I mean, there are changes which will affect urgent 
> needs (e.g, security fixes) and some development (for example, major 
> version updates).
> As for the current plan, there are several links
> http://wiki.openindiana.org/oi/Packaging+projects
> http://wiki.openindiana.org/oi/TODO+list
>
> As for my personal plans, I'd like to look at porting Mate now. As for 
> other tasks, I think we have to remove python2.6 from OI, but to make 
> this real, we have either:
> 1) officially state that /dev is dead and try to merge python2.7 
> changes to illumos-gate,

Just to make sure people understand, that is because not contributing to 
/dev in the first place.

That would require to have some hipster release that can be positively 
upgraded to.

Actually it would be best that hispter releases land in /dev
because hipster's way of landing new packages without announcements, 
plannings and sometime one sentence of description is not what wider 
audience would expect to see in releases.

> 2) invent some decision which would work on OI /dev and old python,

That could be branded zone with 151a9.

I am using hipster on laptop for personal use, lack of alternatives and 
because there needs to be someone who reports bugs and there are always 
many bugs, so many that I am not reporting them on tracker but just on 
IRC, not to make impression everything is full of bugs... (but they get 
fixed gradually as they could, usually by alp alone)

I also think there could be setting what is important and what currently 
is not and work on important things and not on everything one just wants to.
I think it is important to:
-resurrect GUI update,  (python devs welcome)
-maintain GUI installation (python also)
-avoid all core dumps from main applications (testers, bug reporting)
-have hipster recompiled fully
-have OI hipster's own illumos patching process. (maintainers, illumos dev)
-gksudo, gksu does not work,
-revert removal of GUI administration tools
-have pfexec working by default with administrative account (possibly 2 
accounts by fresh install, one with pfexec)
-fix issue of source availability of hipster's packages, because hipster 
does not use OI's own infrastructure for storing source changes and it 
is needed to have sources for every build available on OI servers to 
fulfill licenses, at least in releases. (maintainers)
- standby/resume to make work again (it used to work on hipster till 
first half of 2014). (illumos or hipster issue?)
- last but not least: both Firefox and Thunderbird lost their spellcheck

Some of them are essential (GUI update, GUI install, core apps not core 
dumping, recompliation, source availability, illumos patchig, default 
pfexec for admin account)
and some are not and I am sure everyone can have it's own list but some 
dort of attitude toward releasing something needs to be established.

That could all go before porting Mate and make things with OI /dev to 
hipster update even more complicated.
There are simply more important things then that.

There got to be some hipster release (maybe till the end of the year) 
that people can update from /dev to new /dev made out of hipster.
Landing hipster release in /dev can boost up OI brand and inclusion of 
more people reporting bugs and contributors.

I can suggest people follow hipster for testing and everyday use and 
contribute to changes that lead to stabilization,
not adding more buggy packages ,  therefore, respecting community need 
for releases instead of never ending rolling updates.
Contributors willing to fix listed bugs and make release out of hipster 
that could be upgradeable and land into /dev are welcome.