[OSGeo-Discuss] how did the Paris code sprint go

[Sandro Santilli]

On Mon, Mar 07, 2016 at 07:03:31PM +0000, Hogan, Patrick (ARC-PX) wrote:

> A truly exemplary individual effort for all sprinters, 
> for how great work well precedes the sprint itself, 
> allowing Paul to make the most of this coding event. 

It might be of interest that the bug which kept Paul busy for 70%
of his time was spotted by me 6 weeks before the sprint [1] and
analyzed by me [2] during the sprint. And I also provided a fix
during the sprint [3].

This is not to say that Paul work wasn't needed (he did a great
work modernizing the module inizialization code and backporting
the fix in all branches) but is just to stress out that it doesn't
take being in the same room at the same time to move a project
forward.

Donating some fun to core developers is surely appreciated
(hint: you can pay me a movie theater via bitcoin or paypal)
but I think reliable infrastructure to improve open collaboration
is still the best investiment for a software foundation.

For example, we lost the OSGeo BuildBot instance long time ago
and I'm still missing it [4]. The OSGeo BuildBot infrastructure allowed
any user or packager of every OSGeo project to plug a machine to
the system in order to ensure compatibility of the software with
that specific machine. At the moment, as more and more projects
move to the proprietary-backed "github" infrastructure, most testing
reduced to a specific version of an "ubuntu" system, with some
custom-crafted solutions in each project (Jenkins for PostGIS, Dash
for QGIS, ...).

Maybe we should get all build bots togheter in a room somewhere on
the planet and fill them with beer. It might work :P

[1] https://trac.osgeo.org/postgis/ticket/3429
[2] https://trac.osgeo.org/postgis/ticket/3429#comment:26
[3] https://trac.osgeo.org/postgis/ticket/3429#comment:29
[4] https://trac.osgeo.org/osgeo/ticket/1214

--strk;