[OSGeo-Discuss] Buildbot resurrection, again (was: how did the Paris code sprint go)

[Mateusz Loskot]

On 8 March 2016 at 19:42, Sandro Santilli <strk at keybit.net> wrote:
> On Tue, Mar 08, 2016 at 03:32:54PM +0000, Jody Garnett wrote:
>> I second the idea of build boxes being very important to the well being of
>> projects. I was not aware that OSGeo had been running one, or that it had
>> been taken away.
>
> It was set up by Mateusz, did run wonderfully and got later
> discontinued during some sort of upgrade, I think.

AFAIR, it was combination of resources shortage, migration from
Peer1 to new infrastructure, etc...and the gradual duty take over by
the Travis CI/AppVeyor and alike

> He can surely add some more info. Chances are there's still a
> configuration around, possibly even under revision control.

https://svn.osgeo.org/osgeo/buildbot/

> Right now we have nothing ready to offer to those users that would
> like to donate some of their machines time to ensure PostGIS works
> on their system.

Has there been any poll to see what is potential number of such nodes
willing to connect?

>> Sandro we are setting up priorities for the foundation, I really encourage
>> you to take part and submit a PostGIS build box as an idea. I know the real
>> stop gap is volunteer time, I would hope such a useful idea would attract
>> people.
>
> I recall the OSGeo buildbot was not postgis-specific.
> Mateusz did layout a flexible configuration serving multiple projects.

It was flexible, but still not optimal.
However, IMHO, features offered to required resources ratio would never
beat what FOSS projects have got offered by Travis CI/AppVeyor.

I have already expressed my opinion [1] that I see resurrection of
OSGeo Buildbot a waste of resources.

Even OSGeo hosted GitLab + GitLab CI does not seem resource
efficient, unless majority of OSGeo projects say they would move
from OSgeo SVN (and GitHub too).

No critical mass, nothing to discuss.

[1] http://trac.osgeo.org/osgeo/ticket/1214#comment:3

Best regards,
-- 
Mateusz Loskot, http://mateusz.loskot.net