[Qgis-developer] Rethinking the testing and release procedure of QGIS

[Sandro Santilli]

On Tue, Jul 12, 2011 at 09:20:10AM +0200, Pirmin Kalberer wrote:
> Am Montag, 11. Juli 2011, um 14.06:24 schrieb Ivan Mincik:
> > On 07/11/2011 10:28 AM, Sandro Santilli wrote:
> > > On Sun, Jul 10, 2011 at 10:16:28PM +0200, Ivan Mincik wrote:
> > >>>> Agreed: AFAICT this is a major task: any idea of how much work would
> > >>>> be necessary for reasonable results? An order of magnitude would be
> > >>>> enough for now.
> > >>> 
> > >>> I belive a 30k figure might get things going far enough that developers
> > >>> could have a policy of no functional change and no bugfix w/out a test
> > >>> accompanying would be allowed.
> > >> 
> > >> Sandro, You mean 30 000 EUR?
> > > 
> > > Yes. That'd be 6 man-months. Do you think I'm underestimating it ?
> > 
> > No, no. Just I was not sure if it means that. It is important to know
> > exactly how much money we need to collect.
> 
> It could be dangerous to mix up Sandros estimation with the work for a 
> complete QGIS test suite. QGIS is heavily GUI based which makes writing tests 
> far more complex than writing tests for a pure data processing library like 
> PostGIS e.g. My estimation is 2 to 3 man-years for a test coverage of 80%. 
> Think of topologocal editing, interactive print composer, image comparison
> and so on.

Note that my estimation wasn't for full testcases coverage, but just
for setting up the framework(s) and policies allowing coders to _provide_
testcases in an acceptably simple way, and builders to _run_ tests in
an easy way.

Neither GEOS nor PostGIS has 80% of test coverage, but there's all you need
to add yours, and running them all is as simple as 'make check'.

--strk;

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html