[indiana-discuss] [desktop-discuss] [sw-porters-discuss] contrib and pending repo processes

[John Sonnenschein]

On 18-Nov-08, at 3:08 PM, Shawn Walker wrote:

> Luis de Bethencourt wrote:
>> On Tue, Nov 18, 2008 at 10:46 PM, Nicolas Williams
>> <Nicolas.Williams at sun.com> wrote:
>>> On Tue, Nov 18, 2008 at 02:39:21PM -0800, John Sonnenschein wrote:
>>>> On 18-Nov-08, at 2:21 PM, Nicolas Williams wrote:
>>>>
>>>>> On Tue, Nov 18, 2008 at 02:09:30PM -0800, John Sonnenschein wrote:
>>>>>> That would be acceptable. I'd still prefer the code to be built  
>>>>>> on an
>>>>>> internal machine such that we have an exact record, but I'm  
>>>>>> willing
>>>>>> to
>>>>>> bend this far.
>>>>> I think you'd have to insist that it be [re-]built by "Sun  
>>>>> employees
>>>>> on
>>>>> isolated machines" (certainly not just on SWAN!).
>>>>>
>>>>> For an open community I think that's likely too much.  No?
>>>> Sorry, not Sun internal, but OpenSolaris internal/controlled. If
>>>> Dennis Clarke for example runs the machine and everyone decides  
>>>> that
>>>> he's trusted enough to run the build infrastructure, that's fine  
>>>> too
>>>> IMO, it's more about the controlled build environment.
>>> OK, so you want trusted builders only -- anyone can contribute spec
>>> files, or what have you, but only trusted builders can submit built
>>> pkgs.
>>>
>>> I think that can happen naturally with votes to begin with.  We  
>>> could
>>> start with that and see what happens.
>>>
>>> Alternatively we should hurry up with a spec file-based  
>>> consolidation
>>> and go from there.  That's a *big* change in direction, at least
>>> outwardly, though in a way that's exactly what JDS and pkgfactory  
>>> are
>>> doing, so perhaps not such a big deal.
>>>
>>> Nico
>>> --
>>> _______________________________________________
>>> indiana-discuss mailing list
>>> indiana-discuss at opensolaris.org
>>> http://mail.opensolaris.org/mailman/listinfo/indiana-discuss
>>>
>> I agree with all Nico points. Spec file based submissions built in a
>> controlled machine.
>
> ...where applicable and possible.

This much should be taken as implied, and things that don't fit the  
mold are taken on a case by case basis but It's important to have a  
general set of process not hamstrung by the existence of aberrations